wassname/openrouter-python-sdk-retry-errors
1
0
Fork 0
mirror of https://github.com/wassname/openrouter-python-sdk-retry-errors.git synced 2026-08-18 12:10:34 +08:00
Code Issues Packages Projects Releases Wiki Activity

chore: 🐝 Update SDK - Generate (spec change merged) 0.11.32 (#452)

Browse Source 
Co-authored-by: speakeasybot <bot@speakeasyapi.dev>
Co-authored-by: speakeasy-github[bot] <128539517+speakeasy-github[bot]@users.noreply.github.com>
This commit is contained in:
github-actions[bot]
2026-07-14 21:31:07 +00:00
committed by GitHub
co-authored by speakeasybot speakeasy-github[bot] <128539517+speakeasy-github[bot]@users.noreply.github.com>
parent f07b760c24
commit a3d8d7084d
55 changed files with 473 additions and 334 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.speakeasy/gen.lock
+105 -105
View File
@@ -1,19 +1,19 @@
lockVersion: 2.0.0 lockVersion: 2.0.0
id: c48cf606-fb42-4a45-9c23-8f0555307828 id: c48cf606-fb42-4a45-9c23-8f0555307828
management: management:
docChecksum: 71daa6f89575defb9d76a73e5fb672ae docChecksum: d4956b75e623f552544bf0eee219fb15
docVersion: 1.0.0 docVersion: 1.0.0
speakeasyVersion: 1.787.0 speakeasyVersion: 1.787.0
generationVersion: 2.914.0 generationVersion: 2.914.0
releaseVersion: 0.11.31 releaseVersion: 0.11.32
configChecksum: 96b0d171f179dadfbe9d5efdfff0c4c2 configChecksum: 2def3221df426fb9e1ca1fb865534f8e
repoURL: https://github.com/OpenRouterTeam/python-sdk.git repoURL: https://github.com/OpenRouterTeam/python-sdk.git
installationURL: https://github.com/OpenRouterTeam/python-sdk.git installationURL: https://github.com/OpenRouterTeam/python-sdk.git
published: true published: true
persistentEdits: persistentEdits:
generation_id: 056b9ee4-a227-4cc6-b7e0-3c97d424d043 generation_id: c6eafc4f-1723-4a50-b64a-768ca4b13a2f
pristine_commit_hash: 9daff37b2becb50eb2fe167e87e1149e4213b27e pristine_commit_hash: 2be2ba16cd458078a20e8dff747aed3a7eeba223
pristine_tree_hash: 26d42d8bfd711f5fb23141b6ab241881fd7137a9 pristine_tree_hash: d075876a76e82a3c03a6789df86ec9dfede361bb
features: features:
python: python:
acceptHeaders: 3.0.0 acceptHeaders: 3.0.0
@@ -234,8 +234,8 @@ trackedFiles:
pristine_git_object: 44deb61a813fde8473d34e0a14ccaef643467e0d pristine_git_object: 44deb61a813fde8473d34e0a14ccaef643467e0d
docs/components/anthropiccachecontroldirective.mdx: docs/components/anthropiccachecontroldirective.mdx:
id: 5671704f9917 id: 5671704f9917
last_write_checksum: sha1:03d0b8bed7d5dd448141a09c4515594bcc7ef3f2 last_write_checksum: sha1:a31ffcd83295557f5a070d96873408214676a2ea
pristine_git_object: 5f755985a972da22e96483e3807cb2029ba18989 pristine_git_object: 12b753fe8b98723cd210a3dac91402a026b1dfac
docs/components/anthropiccachecontroldirectivetype.mdx: docs/components/anthropiccachecontroldirectivetype.mdx:
id: f5d844d35b58 id: f5d844d35b58
last_write_checksum: sha1:9e513498c44431258e61e8fb25537d8f6c5d4f7d last_write_checksum: sha1:9e513498c44431258e61e8fb25537d8f6c5d4f7d
@@ -298,8 +298,8 @@ trackedFiles:
pristine_git_object: 2b04c2483fceacf8300bb872ea0eed600970d160 pristine_git_object: 2b04c2483fceacf8300bb872ea0eed600970d160
docs/components/anthropicdocumentblockparam.mdx: docs/components/anthropicdocumentblockparam.mdx:
id: ece98dcf04f6 id: ece98dcf04f6
last_write_checksum: sha1:5a05c624a5a134bedac12d687f51621448627daf last_write_checksum: sha1:42f908ffe281fd4272a24d7eb874f76fbee47da9
pristine_git_object: 5a24251490513dca5c64fa3f8b29c80eafdcd331 pristine_git_object: 1ed7b7cd83e412bdc9945f00a875bc681e0c96ee
docs/components/anthropicdocumentblockparamcitations.mdx: docs/components/anthropicdocumentblockparamcitations.mdx:
id: 474518ae0167 id: 474518ae0167
last_write_checksum: sha1:798372e77425fd7127a318c75247ce44e6f1dbfd last_write_checksum: sha1:798372e77425fd7127a318c75247ce44e6f1dbfd
@@ -326,8 +326,8 @@ trackedFiles:
pristine_git_object: c86bc52963eca6a9de705f418c56d17a4bf8ebc6 pristine_git_object: c86bc52963eca6a9de705f418c56d17a4bf8ebc6
docs/components/anthropicimageblockparam.mdx: docs/components/anthropicimageblockparam.mdx:
id: a85ff43edd3a id: a85ff43edd3a
last_write_checksum: sha1:b2324840f7be26004674f2a59f5d66122b75437a last_write_checksum: sha1:32a650d11a80f74059407954fa86eba2247fdf03
pristine_git_object: 2d95dd47a3b3f4551cd4a148f4936057a40587f5 pristine_git_object: 9be25eff134f6283b95213ba56fa2ee21c493a39
docs/components/anthropicimageblockparamsource.mdx: docs/components/anthropicimageblockparamsource.mdx:
id: 82f05c79d5cb id: 82f05c79d5cb
last_write_checksum: sha1:0939a26da32a1d150a58d6988b9ef8bbd49bacf4 last_write_checksum: sha1:0939a26da32a1d150a58d6988b9ef8bbd49bacf4
@@ -382,8 +382,8 @@ trackedFiles:
pristine_git_object: a538343b4f3fca39a44d3f208e702c482978c4f8 pristine_git_object: a538343b4f3fca39a44d3f208e702c482978c4f8
docs/components/anthropicsearchresultblockparam.mdx: docs/components/anthropicsearchresultblockparam.mdx:
id: 5c62b24b2c58 id: 5c62b24b2c58
last_write_checksum: sha1:a8acf36f4eb32d7d01a906f889f9c76e36786f24 last_write_checksum: sha1:b6f01f3da19d4a430e66f5c47e90f31aa3fb0e29
pristine_git_object: 517785e7ab5c58a8e41292572d2d581a0a0e3df7 pristine_git_object: 89ae1afd55149a2e641d5330189fd5213aa7f05b
docs/components/anthropicsearchresultblockparamcitations.mdx: docs/components/anthropicsearchresultblockparamcitations.mdx:
id: e6a0db284ce9 id: e6a0db284ce9
last_write_checksum: sha1:20a8315824e11129d12d4d5c441e9fde3691a1f5 last_write_checksum: sha1:20a8315824e11129d12d4d5c441e9fde3691a1f5
@@ -398,8 +398,8 @@ trackedFiles:
pristine_git_object: 6f1e9ce880a0c3173ce7af4d14c8508b686b8ebc pristine_git_object: 6f1e9ce880a0c3173ce7af4d14c8508b686b8ebc
docs/components/anthropictextblockparam.mdx: docs/components/anthropictextblockparam.mdx:
id: bded5d409e91 id: bded5d409e91
last_write_checksum: sha1:5c8b4dafc14fafa615c888b9cd207bac55ebb585 last_write_checksum: sha1:ab2ebb14dd5e54bbadf4db4aa919d1b380b844b5
pristine_git_object: 6d00495d81a02e202e2ae1346439d734a75df8f0 pristine_git_object: 502d15e612cfe8df8b28ba89855683eee5f0ce4b
docs/components/anthropictextblockparamtype.mdx: docs/components/anthropictextblockparamtype.mdx:
id: 7f03a13234b7 id: 7f03a13234b7
last_write_checksum: sha1:12c7718c5c0da4561914ed95dd6853873568bd05 last_write_checksum: sha1:12c7718c5c0da4561914ed95dd6853873568bd05
@@ -758,8 +758,8 @@ trackedFiles:
pristine_git_object: a7472d01b6de43fedfa76a9e608d67202a7ec529 pristine_git_object: a7472d01b6de43fedfa76a9e608d67202a7ec529
docs/components/caching.mdx: docs/components/caching.mdx:
id: edc8a0f8f37f id: edc8a0f8f37f
last_write_checksum: sha1:b4e6619f389c1440555ec567fbdcddbc9976e63e last_write_checksum: sha1:cce3835ca76b640b591de1307422aa4e391cde69
pristine_git_object: d9c719912c5e82aa2bf9b02c54fb03c8d448ef83 pristine_git_object: 0d7013f54bed0390239fa49074a48d5b21711a28
docs/components/capabilitydescriptor.mdx: docs/components/capabilitydescriptor.mdx:
id: 34ac0e581d79 id: 34ac0e581d79
last_write_checksum: sha1:62ee9ccab6cec4572383cef654997ab4106d4676 last_write_checksum: sha1:62ee9ccab6cec4572383cef654997ab4106d4676
@@ -810,8 +810,8 @@ trackedFiles:
pristine_git_object: fdb10076d12248abeacfae19a1bd647d4a838eaa pristine_git_object: fdb10076d12248abeacfae19a1bd647d4a838eaa
docs/components/chatcontentcachecontrol.mdx: docs/components/chatcontentcachecontrol.mdx:
id: 215e2356f770 id: 215e2356f770
last_write_checksum: sha1:cf9c5e64dcedcccbab8d79499bea7bf76e86b1b9 last_write_checksum: sha1:96bb8d2d2c51fc1b5208120f84ca175ddcd0346a
pristine_git_object: 250fd47f76c936c2e979b55caa6d4fb17b4a26ef pristine_git_object: 6f1d2f8e0e284e16eeff1f0dbd03166260f3ac94
docs/components/chatcontentcachecontroltype.mdx: docs/components/chatcontentcachecontroltype.mdx:
id: 93c1e9a9146e id: 93c1e9a9146e
last_write_checksum: sha1:febb0b8575ef8fc009f71312d8fed2b596cf6d28 last_write_checksum: sha1:febb0b8575ef8fc009f71312d8fed2b596cf6d28
@@ -846,8 +846,8 @@ trackedFiles:
pristine_git_object: 92c9c5c2ce23807bac9756e55803057921c34f24 pristine_git_object: 92c9c5c2ce23807bac9756e55803057921c34f24
docs/components/chatcontenttext.mdx: docs/components/chatcontenttext.mdx:
id: a2436d6c5c1c id: a2436d6c5c1c
last_write_checksum: sha1:dfda97beb6d5cdb859b9207798e34ed7ed4691f8 last_write_checksum: sha1:bf50de67c2c4b71ef14a92a53507b1c28fa1cc5b
pristine_git_object: a59dafe92bfb650c89244eaa89f00690fd8282d3 pristine_git_object: b31006bd9cd42c1afa9f936eb085af94ae786f4d
docs/components/chatcontenttexttype.mdx: docs/components/chatcontenttexttype.mdx:
id: eb29ddb48a88 id: eb29ddb48a88
last_write_checksum: sha1:154a7cb1f0dcdc0ff60b5fcfe1ce5e86e08bb72f last_write_checksum: sha1:154a7cb1f0dcdc0ff60b5fcfe1ce5e86e08bb72f
@@ -930,8 +930,8 @@ trackedFiles:
pristine_git_object: bc9f2fa0780e97e0d13a2c10af27a5f3e666c0f5 pristine_git_object: bc9f2fa0780e97e0d13a2c10af27a5f3e666c0f5
docs/components/chatfunctiontoolfunction.mdx: docs/components/chatfunctiontoolfunction.mdx:
id: b9f813a5b0c9 id: b9f813a5b0c9
last_write_checksum: sha1:1daf25e4f31b7013fcbcd43f86678ce1548c8423 last_write_checksum: sha1:649b1fe5d94d5613d6e902d5d96762be33e648eb
pristine_git_object: 7619eec425176b875101f822b806ff98a5d991c9 pristine_git_object: a6cf73c17fd5cf6752d09f7eb25a13fb8d9d4a02
docs/components/chatfunctiontoolfunctionfunction.mdx: docs/components/chatfunctiontoolfunctionfunction.mdx:
id: 59c7a37a65c9 id: 59c7a37a65c9
last_write_checksum: sha1:671313b618e2e45e37f94494b8bd88822aa4220a last_write_checksum: sha1:671313b618e2e45e37f94494b8bd88822aa4220a
@@ -966,8 +966,8 @@ trackedFiles:
pristine_git_object: 220be03bd674a257c2064c0455104959b3974a35 pristine_git_object: 220be03bd674a257c2064c0455104959b3974a35
docs/components/chatrequest.mdx: docs/components/chatrequest.mdx:
id: 055812be14d8 id: 055812be14d8
last_write_checksum: sha1:77fcd137afe54399a3eba5f08cf6e0ffda8cc242 last_write_checksum: sha1:8803f814796b25653c13db7cd2e167720d615499
pristine_git_object: 1cc4f0f490833b0defa14840b69ea02f90820770 pristine_git_object: e831058c3a681c3f9e8896137e06d82d3c416a5a
docs/components/chatrequesteffort.mdx: docs/components/chatrequesteffort.mdx:
id: 703c73252ebd id: 703c73252ebd
last_write_checksum: sha1:243510f188071a3f317dc06a07fbcfcedd3cf821 last_write_checksum: sha1:243510f188071a3f317dc06a07fbcfcedd3cf821
@@ -1242,8 +1242,8 @@ trackedFiles:
pristine_git_object: 9a8e9eaadebb36e2836d73c4ed1867c5f1e564e7 pristine_git_object: 9a8e9eaadebb36e2836d73c4ed1867c5f1e564e7
docs/components/contentcompaction.mdx: docs/components/contentcompaction.mdx:
id: 1923f9db3cff id: 1923f9db3cff
last_write_checksum: sha1:59526732df9d4b5a69f3a79d1210493dff8d8898 last_write_checksum: sha1:db0ae20faae38b060e586c80655b79849bb4e18b
pristine_git_object: 39acf3a7dc4de67d2eb07a4e21e33e6cfdd2091f pristine_git_object: 87f9c8d7936595373ebeae59a265705db8df2721
docs/components/contentencryptedcontent.mdx: docs/components/contentencryptedcontent.mdx:
id: 0dc55809c394 id: 0dc55809c394
last_write_checksum: sha1:97675b67fb3cdb6c892bbe01d2cd9e6bab5e4dec last_write_checksum: sha1:97675b67fb3cdb6c892bbe01d2cd9e6bab5e4dec
@@ -1354,8 +1354,8 @@ trackedFiles:
pristine_git_object: f9c4eeefa07fcc73839934dd6b3d403fe7539da3 pristine_git_object: f9c4eeefa07fcc73839934dd6b3d403fe7539da3
docs/components/contentservertooluse.mdx: docs/components/contentservertooluse.mdx:
id: 7b676db3a465 id: 7b676db3a465
last_write_checksum: sha1:65f0643d0a7af0e2ef8b4f2dc9a336b5a34b135c last_write_checksum: sha1:b49ee690460e4de02996066b31e91bb1f4eb8400
pristine_git_object: 527a1c1ae4e8df68528db2404024b67348ab4b6c pristine_git_object: 86c112ede3a14c30546d1fb00f6526a5f854ced0
docs/components/contentthinking.mdx: docs/components/contentthinking.mdx:
id: c23024acf25a id: c23024acf25a
last_write_checksum: sha1:f3a989b8eb7a38f2705324263796f49a7e1722b6 last_write_checksum: sha1:f3a989b8eb7a38f2705324263796f49a7e1722b6
@@ -1366,16 +1366,16 @@ trackedFiles:
pristine_git_object: f07b01d7d44cc717b67e5648283c659fc02c72fa pristine_git_object: f07b01d7d44cc717b67e5648283c659fc02c72fa
docs/components/contenttoolresult.mdx: docs/components/contenttoolresult.mdx:
id: a9a083b8900f id: a9a083b8900f
last_write_checksum: sha1:1a0bb46e6c27779f1f88da7eb87d0ae04acd62af last_write_checksum: sha1:72b603d3efa32360cf6e8b6c718dd46718b0e98b
pristine_git_object: 4abde669843735aeb36b10c6a5992f8d6509cd40 pristine_git_object: f83588a9c24d3172b8f8fd131cdc5aea073d0088
docs/components/contenttooluse.mdx: docs/components/contenttooluse.mdx:
id: ad7fb52d642b id: ad7fb52d642b
last_write_checksum: sha1:47f463246904b67a444c6072b1d927c1c6c223de last_write_checksum: sha1:96dda7c7f895597f5f60c57b69936f77e14fc8ff
pristine_git_object: e29dae101a0e629b6cfa1473fc22d732e646ef55 pristine_git_object: 7149ea56376799b7d485254b63acdda00e9713a4
docs/components/contentwebsearchtoolresult.mdx: docs/components/contentwebsearchtoolresult.mdx:
id: d2401cf2f6b9 id: d2401cf2f6b9
last_write_checksum: sha1:b24e06b9ba722674a7fc4116eba838657914f166 last_write_checksum: sha1:d7ecfe7df3b1e89151edb59cd1525aab997b28b6
pristine_git_object: a2664d14e56d9f5ada405cee06d7f745e1a5a47b pristine_git_object: 9a9180dcba1e86c7a26c5396f39bc4dafe160e96
docs/components/contentwebsearchtoolresulterror.mdx: docs/components/contentwebsearchtoolresulterror.mdx:
id: e53d83e87e2b id: e53d83e87e2b
last_write_checksum: sha1:22dd4ec3f699feb7d8f22abff79623b427292e35 last_write_checksum: sha1:22dd4ec3f699feb7d8f22abff79623b427292e35
@@ -2010,8 +2010,8 @@ trackedFiles:
pristine_git_object: 3c29bc4bd3bdf7037b9c3aa894180b8e4cae6098 pristine_git_object: 3c29bc4bd3bdf7037b9c3aa894180b8e4cae6098
docs/components/fusionservertoolconfig.mdx: docs/components/fusionservertoolconfig.mdx:
id: 87976811b0c3 id: 87976811b0c3
last_write_checksum: sha1:99198fd63b66a6f465c444856fcfb35110f269fb last_write_checksum: sha1:267106d40b0f8e9a7601945753f7934b34084183
pristine_git_object: ec3ad69133064ba5a6911474a16a974a26963255 pristine_git_object: 1247ee7797392ca5e7778a47b42932a97cc16203
docs/components/fusionservertoolconfigeffort.mdx: docs/components/fusionservertoolconfigeffort.mdx:
id: f5ce0fbc94aa id: f5ce0fbc94aa
last_write_checksum: sha1:4ce09a069b4eeae41ac1531027cf97dc4139927f last_write_checksum: sha1:4ce09a069b4eeae41ac1531027cf97dc4139927f
@@ -2498,8 +2498,8 @@ trackedFiles:
pristine_git_object: 2e56403330ecae1d43fd2e4d23976dfe5330c38d pristine_git_object: 2e56403330ecae1d43fd2e4d23976dfe5330c38d
docs/components/inputtext.mdx: docs/components/inputtext.mdx:
id: c8a97322a50a id: c8a97322a50a
last_write_checksum: sha1:58653e40ea0a33629dda40d2a11c8d21a75b0467 last_write_checksum: sha1:6f92abbb193a5d21cc129df4c5804b3f9161322d
pristine_git_object: ece32d6edb601578cdb56fc8b4a60e40bc089f22 pristine_git_object: 5908413083a34d8bdd34ed916610dc5eee0d6564
docs/components/inputtexttype.mdx: docs/components/inputtexttype.mdx:
id: 91f958925357 id: 91f958925357
last_write_checksum: sha1:42a8d9a6d3506e02f6024592691292b7ac255075 last_write_checksum: sha1:42a8d9a6d3506e02f6024592691292b7ac255075
@@ -2770,8 +2770,8 @@ trackedFiles:
pristine_git_object: cff8f3938a8bd5699d1111a6b1a926518d868da5 pristine_git_object: cff8f3938a8bd5699d1111a6b1a926518d868da5
docs/components/messagesrequest.mdx: docs/components/messagesrequest.mdx:
id: 762dd6b894c8 id: 762dd6b894c8
last_write_checksum: sha1:8e06b3369eb4dee0ab5fee504e444cb5be1f2f3c last_write_checksum: sha1:a2837f5f14f370453c77fd6962dea4fa9fa43999
pristine_git_object: 2b25d420887ade7549b1848592fb24ed622b54f1 pristine_git_object: 89b9780ced7cfd41a0f2b0296232df38e71b84ee
docs/components/messagesrequestmetadata.mdx: docs/components/messagesrequestmetadata.mdx:
id: facdcaeb2576 id: facdcaeb2576
last_write_checksum: sha1:42bf112bc4f6d259c589906fd5f67c5342ab4c2d last_write_checksum: sha1:42bf112bc4f6d259c589906fd5f67c5342ab4c2d
@@ -3898,8 +3898,8 @@ trackedFiles:
pristine_git_object: 5fb1b765b2b0e46d40042f341bbcc91aa0d92762 pristine_git_object: 5fb1b765b2b0e46d40042f341bbcc91aa0d92762
docs/components/promptcachebreakpoint.mdx: docs/components/promptcachebreakpoint.mdx:
id: c135956dc623 id: c135956dc623
last_write_checksum: sha1:1b0c9884eb9b6fe1d4fdf822c75cb566ead67e07 last_write_checksum: sha1:6df345ea99da1fdcae22895faa2ee36339ba7332
pristine_git_object: 1d1ff6162c98c1639ea8e2e4dfa7f6d5cd7d3fcd pristine_git_object: 313ab84d0db5d181f8dea91208a4b5127f43477b
docs/components/promptcachebreakpointmode.mdx: docs/components/promptcachebreakpointmode.mdx:
id: 05abe4e54869 id: 05abe4e54869
last_write_checksum: sha1:f7ca9f8ff280c1fa750609f4aa13b0706d1dd7cd last_write_checksum: sha1:f7ca9f8ff280c1fa750609f4aa13b0706d1dd7cd
@@ -4246,8 +4246,8 @@ trackedFiles:
pristine_git_object: 996cffee821435729436d7f29bce4333f4900382 pristine_git_object: 996cffee821435729436d7f29bce4333f4900382
docs/components/responsesrequest.mdx: docs/components/responsesrequest.mdx:
id: 0dbcef40a4b1 id: 0dbcef40a4b1
last_write_checksum: sha1:462f684189fca9eefc0809564ab964d89025724d last_write_checksum: sha1:51003cb482d6de367ebe0c4e2691bf41be8d384c
pristine_git_object: 371d94f398a631b4146a2a76b46dbd257637caba pristine_git_object: 3090f61cfb772d83d2ac230fb571ef3e048e423e
docs/components/responsesrequestplugin.mdx: docs/components/responsesrequestplugin.mdx:
id: 4d550c41c85f id: 4d550c41c85f
last_write_checksum: sha1:ab9d1e6c062cd51fc976da6ccf393e9974171f7b last_write_checksum: sha1:ab9d1e6c062cd51fc976da6ccf393e9974171f7b
@@ -4690,12 +4690,12 @@ trackedFiles:
pristine_git_object: 54a7627079b16a50865cf6f2224abe36870c95fa pristine_git_object: 54a7627079b16a50865cf6f2224abe36870c95fa
docs/components/tooladvisor20260301.mdx: docs/components/tooladvisor20260301.mdx:
id: df2858dddd07 id: df2858dddd07
last_write_checksum: sha1:94a6e30c4bc7ff7771b99d3f09c87604423b2847 last_write_checksum: sha1:2dd846c19c5dabfeba4fe838ca60bd495c31e689
pristine_git_object: ee1f63bc170b35e5fb8da3390818160a2b4b93c6 pristine_git_object: a00186b54ba47296aed2eb1098928f3b4a565ceb
docs/components/toolbash20250124.mdx: docs/components/toolbash20250124.mdx:
id: acc36fba0c55 id: acc36fba0c55
last_write_checksum: sha1:b0e3ad180f48dba955ae20481ae6ef55ace6b18f last_write_checksum: sha1:b899d22b5559c09fcaa14c5cd3ff18f914ad34d2
pristine_git_object: d9f17ef27ea42b2251c85ffdd230af45f7293c75 pristine_git_object: d3a3ea6cfbc5118b9adad9c1a8e543e02fd2870f
docs/components/toolcallstatus.mdx: docs/components/toolcallstatus.mdx:
id: e79e5ffbe798 id: e79e5ffbe798
last_write_checksum: sha1:338789050d562f0ac596d6a94783307e27a634b2 last_write_checksum: sha1:338789050d562f0ac596d6a94783307e27a634b2
@@ -4734,12 +4734,12 @@ trackedFiles:
pristine_git_object: 6817216b64de42e834ab68a072287701e621daa8 pristine_git_object: 6817216b64de42e834ab68a072287701e621daa8
docs/components/toolcustom.mdx: docs/components/toolcustom.mdx:
id: b3ea2dbc7af5 id: b3ea2dbc7af5
last_write_checksum: sha1:dd0734eb86cbbba753632268613a2f2ebc8e8f78 last_write_checksum: sha1:f0fdef1bc45a22946f6026d202499b63ff241711
pristine_git_object: 57073f649edd84b26f2395c46aef55c86e31ed4c pristine_git_object: ce9be3cd6fbb03e8d15866baefa9f38c1398a91c
docs/components/tooltexteditor20250124.mdx: docs/components/tooltexteditor20250124.mdx:
id: 58d2a6f35f9f id: 58d2a6f35f9f
last_write_checksum: sha1:bbc36d5c8d48b8971fed82553489759ca2379c0d last_write_checksum: sha1:c785aecfc38109c995d4d2f825de0f145d3f626a
pristine_git_object: 0fe06a930a6a67a5a007185f892422f232389d3a pristine_git_object: a7f1a2731ba8ab1deea97b51e63372e70e3afa0e
docs/components/tooltypecustom.mdx: docs/components/tooltypecustom.mdx:
id: ac3c7f5f5fa6 id: ac3c7f5f5fa6
last_write_checksum: sha1:bfbecb3b4e70cb8a4aa0379246f4ffbaae71b341 last_write_checksum: sha1:bfbecb3b4e70cb8a4aa0379246f4ffbaae71b341
@@ -4750,12 +4750,12 @@ trackedFiles:
pristine_git_object: 48e6f8afc2a41ccac4be220dd5d5d475a4d36b71 pristine_git_object: 48e6f8afc2a41ccac4be220dd5d5d475a4d36b71
docs/components/toolwebsearch20250305.mdx: docs/components/toolwebsearch20250305.mdx:
id: e583eb0b3bd1 id: e583eb0b3bd1
last_write_checksum: sha1:59297ba6e78d528a5a6cc0bcbefd1301c21af308 last_write_checksum: sha1:a750dd139f327bf53d4cde02e6b897afa274475b
pristine_git_object: 33510135de4236e72077c4acb6c8a8c65fe5f647 pristine_git_object: 36a096412525a09d1903488c804648047782b8be
docs/components/toolwebsearch20260209.mdx: docs/components/toolwebsearch20260209.mdx:
id: dfc99c1c297c id: dfc99c1c297c
last_write_checksum: sha1:e3adde8067e2e5e5504165e7b8ebf1ca743d8352 last_write_checksum: sha1:8237527944c9d8670c8c323ce8bec709602d8322
pristine_git_object: a70393e56b803a455f0ea3e82d9270ee84fed5fe pristine_git_object: 062b3155e1ffd04dae559ccc126ff8b922616aee
docs/components/toomanyrequestsresponseerrordata.mdx: docs/components/toomanyrequestsresponseerrordata.mdx:
id: 282b92f982c7 id: 282b92f982c7
last_write_checksum: sha1:976995e37e897805df9f3f50c0aae39e41447822 last_write_checksum: sha1:976995e37e897805df9f3f50c0aae39e41447822
@@ -6574,8 +6574,8 @@ trackedFiles:
pristine_git_object: 513bb72efe9f74672b1b60d6b45dc0aca282ddf6 pristine_git_object: 513bb72efe9f74672b1b60d6b45dc0aca282ddf6
docs/sdks/chat/README.mdx: docs/sdks/chat/README.mdx:
id: 1dd859c23fe1 id: 1dd859c23fe1
last_write_checksum: sha1:b6c104dc6b0c3461a2264a95d886f860e7277780 last_write_checksum: sha1:afeaa6d1195ec68d845aa76038b47498826d48a3
pristine_git_object: 390ca7cf2ff70cf7a4273cd1c8c13cb58eb32971 pristine_git_object: 70aca7d4ce8dd7f8356abc4d836e8bdc236e090f
docs/sdks/classifications/README.mdx: docs/sdks/classifications/README.mdx:
id: 4786130ef02a id: 4786130ef02a
last_write_checksum: sha1:87ea3518b4147ae7833b01da7be92d41a4b49df7 last_write_checksum: sha1:87ea3518b4147ae7833b01da7be92d41a4b49df7
@@ -6630,8 +6630,8 @@ trackedFiles:
pristine_git_object: 32a291cb837977e7cdd475c664e5d475660b2ba3 pristine_git_object: 32a291cb837977e7cdd475c664e5d475660b2ba3
docs/sdks/presets/README.mdx: docs/sdks/presets/README.mdx:
id: eeb505928d52 id: eeb505928d52
last_write_checksum: sha1:0c1569e8db293c65840433adda49310960a591d1 last_write_checksum: sha1:77d23418817e3a4e1d7135bdbd21a61812c960ee
pristine_git_object: ef18762b9c9c42a26bf61c24af1607d7ce19c01a pristine_git_object: 82e78fb13dc1c886ea37068b24ad1e5c786b6052
docs/sdks/providers/README.mdx: docs/sdks/providers/README.mdx:
id: 5ea9af111324 id: 5ea9af111324
last_write_checksum: sha1:570b55655ee7602d3807c3218e450741f634eed6 last_write_checksum: sha1:570b55655ee7602d3807c3218e450741f634eed6
@@ -6642,8 +6642,8 @@ trackedFiles:
pristine_git_object: cd1b341b1b7c0284b9ba9b102bd8020e0d7ce707 pristine_git_object: cd1b341b1b7c0284b9ba9b102bd8020e0d7ce707
docs/sdks/responses/README.mdx: docs/sdks/responses/README.mdx:
id: abab319e080e id: abab319e080e
last_write_checksum: sha1:69540ef8f6a556c86c4ba8a6ebd4ed03e36ee927 last_write_checksum: sha1:2dbf6b499e85a4605d79d756e043f7bfbc633087
pristine_git_object: bbd2952598bc9e67b6080bf8ee53d7f6ac838d9c pristine_git_object: 708c8a1312812b7d552b5e970f0fb8c27f61b4df
docs/sdks/stt/README.mdx: docs/sdks/stt/README.mdx:
id: 190b0dc9a5d1 id: 190b0dc9a5d1
last_write_checksum: sha1:a8e68ac2b0bb8640e01d962177b42ab47b24898b last_write_checksum: sha1:a8e68ac2b0bb8640e01d962177b42ab47b24898b
@@ -6666,8 +6666,8 @@ trackedFiles:
pristine_git_object: 3e38f1a929f7d6b1d6de74604aa87e3d8f010544 pristine_git_object: 3e38f1a929f7d6b1d6de74604aa87e3d8f010544
pyproject.toml: pyproject.toml:
id: 5d07e7d72637 id: 5d07e7d72637
last_write_checksum: sha1:836c2f803d896f76e749fcfbae9070ea43880710 last_write_checksum: sha1:bd90540d9cd6c97085fa5fd5f08dbbbfa2899391
pristine_git_object: cc31c2921b9b60ab306c0e306efc843bf847ff03 pristine_git_object: f72d3afab559e9db67b290ca5e6337adfcda58c9
scripts/prepare_readme.py: scripts/prepare_readme.py:
id: e0c5957a6035 id: e0c5957a6035
last_write_checksum: sha1:77f44b60b98bc126557ec27391f91dfba764bb54 last_write_checksum: sha1:77f44b60b98bc126557ec27391f91dfba764bb54
@@ -6694,8 +6694,8 @@ trackedFiles:
pristine_git_object: 86713cfea633e09d33b3d4e65281071fe20e6137 pristine_git_object: 86713cfea633e09d33b3d4e65281071fe20e6137
src/openrouter/_version.py: src/openrouter/_version.py:
id: d8d15ad6c586 id: d8d15ad6c586
last_write_checksum: sha1:94d5d591203cb4700f3decbcd7e864a271821999 last_write_checksum: sha1:cffbb1cf67568b37538f51f21b858e0242adb9cf
pristine_git_object: e72721452a9f36dce4f645bd1d995077867882a8 pristine_git_object: 44f59a476730ba6703b7d5e088bbd3a81e21b362
src/openrouter/analytics.py: src/openrouter/analytics.py:
id: cb406b5aaabb id: cb406b5aaabb
last_write_checksum: sha1:9e709b71dd0611056dc0cec6150b578defe32841 last_write_checksum: sha1:9e709b71dd0611056dc0cec6150b578defe32841
@@ -6726,8 +6726,8 @@ trackedFiles:
pristine_git_object: 08b177b0b39e7f3581f1a5da13e245c3daddf9d5 pristine_git_object: 08b177b0b39e7f3581f1a5da13e245c3daddf9d5
src/openrouter/chat.py: src/openrouter/chat.py:
id: 723fdce15c1d id: 723fdce15c1d
last_write_checksum: sha1:79e9c02b4a16cae3c202b5ae0bfd6cd1d0d6f89b last_write_checksum: sha1:2afda128e7af2900cfeeaf79cc5a52122da8c3a2
pristine_git_object: ab93b1be035fe591f6c8fac84b57ecad89ff5b1a pristine_git_object: f14c3b5188d6dfdefcc28aa685daa24276ae2990
src/openrouter/classifications.py: src/openrouter/classifications.py:
id: 4f2efc24b95b id: 4f2efc24b95b
last_write_checksum: sha1:3ab6fcd9ef679cd2149ec8dfbb4a86109cd152ad last_write_checksum: sha1:3ab6fcd9ef679cd2149ec8dfbb4a86109cd152ad
@@ -6794,8 +6794,8 @@ trackedFiles:
pristine_git_object: 870c2e75705a43762b7aa2797e5b0cde4fe40d87 pristine_git_object: 870c2e75705a43762b7aa2797e5b0cde4fe40d87
src/openrouter/components/anthropiccachecontroldirective.py: src/openrouter/components/anthropiccachecontroldirective.py:
id: ffa72802cd82 id: ffa72802cd82
last_write_checksum: sha1:c73ec323b8d1e650f51d91f00b824804e4b9f6c3 last_write_checksum: sha1:2c92e5081aa549627b961c97cfe0ba8d672a567c
pristine_git_object: a9292b2d0dc104b016f7bc2600275d87fbd9ecac pristine_git_object: 05b36d10c1f772110394071a9c90b1d9c31bf2c2
src/openrouter/components/anthropiccachecontrolttl.py: src/openrouter/components/anthropiccachecontrolttl.py:
id: 2ec3e3546afe id: 2ec3e3546afe
last_write_checksum: sha1:338e193bd6929eb6ed69da9bdb4e608e5db6121b last_write_checksum: sha1:338e193bd6929eb6ed69da9bdb4e608e5db6121b
@@ -6830,16 +6830,16 @@ trackedFiles:
pristine_git_object: 50ad24f32ec87623c006381c477c46f44d84c753 pristine_git_object: 50ad24f32ec87623c006381c477c46f44d84c753
src/openrouter/components/anthropicdocumentblockparam.py: src/openrouter/components/anthropicdocumentblockparam.py:
id: 2c3093530540 id: 2c3093530540
last_write_checksum: sha1:904f1546a7a69df231395f33cb5b1b71ae0c2611 last_write_checksum: sha1:43f9b964f383ea77663e47c9a1e1cd76b7457d2c
pristine_git_object: 686764bb6991204328c5017ab636e3d38a4f2f09 pristine_git_object: be2d48fefcffa05ca80bfa647f77fafe60872a7d
src/openrouter/components/anthropicfiledocumentsource.py: src/openrouter/components/anthropicfiledocumentsource.py:
id: d824b52285d7 id: d824b52285d7
last_write_checksum: sha1:5d3c89e2d60df0116aed28d9a2274cf61536af41 last_write_checksum: sha1:5d3c89e2d60df0116aed28d9a2274cf61536af41
pristine_git_object: 06ab054ee7227cb2186c19c108145d8e19646044 pristine_git_object: 06ab054ee7227cb2186c19c108145d8e19646044
src/openrouter/components/anthropicimageblockparam.py: src/openrouter/components/anthropicimageblockparam.py:
id: 71cab07562d1 id: 71cab07562d1
last_write_checksum: sha1:7077151d8ef411a85bc3cdafb17ae7969c8bf7cb last_write_checksum: sha1:fbdb3a017bc51cb2848933aca87023159951a2aa
pristine_git_object: 979d0464c4ef00aba59e1431a2bbd3568c411246 pristine_git_object: 7b5effbbab60e15e1daa06ad703f445d3dde96b6
src/openrouter/components/anthropicimagemimetype.py: src/openrouter/components/anthropicimagemimetype.py:
id: 2434bf06fc4f id: 2434bf06fc4f
last_write_checksum: sha1:2016e4ef5d678a2af645d50cb02e022c3918d494 last_write_checksum: sha1:2016e4ef5d678a2af645d50cb02e022c3918d494
@@ -6866,16 +6866,16 @@ trackedFiles:
pristine_git_object: d16f6dc9a2449185ef45731cbcd0a78c266745ef pristine_git_object: d16f6dc9a2449185ef45731cbcd0a78c266745ef
src/openrouter/components/anthropicsearchresultblockparam.py: src/openrouter/components/anthropicsearchresultblockparam.py:
id: bccb7c6b3128 id: bccb7c6b3128
last_write_checksum: sha1:aa8fcc04743f07cf8a72b3cbfaae212524101700 last_write_checksum: sha1:8d073fef507a39b53421384ff9a5914fd9438493
pristine_git_object: 5925732967541b0d598ac249f372e5b18f975b0d pristine_git_object: 14967a292ebe05517a28c98b1e241bd2fd03a572
src/openrouter/components/anthropicspeed.py: src/openrouter/components/anthropicspeed.py:
id: 7e39c29f3d89 id: 7e39c29f3d89
last_write_checksum: sha1:19e37bf7ec082d89258f44674405811cc8a010ad last_write_checksum: sha1:19e37bf7ec082d89258f44674405811cc8a010ad
pristine_git_object: 4f9c2bc739541b12c381e3d6265184af879cc1ff pristine_git_object: 4f9c2bc739541b12c381e3d6265184af879cc1ff
src/openrouter/components/anthropictextblockparam.py: src/openrouter/components/anthropictextblockparam.py:
id: f387daeeb940 id: f387daeeb940
last_write_checksum: sha1:266f0b2002e313484812c69d4b85a7d84aeb0c2b last_write_checksum: sha1:d07d4bef39b0f763604a58604da6aa4cb7979590
pristine_git_object: a6a70aa6825c515fb5c9c1358d011314df010b2d pristine_git_object: 31a1df2ee27149c8862fb8709f4898ef0535360f
src/openrouter/components/anthropicthinkingdisplay.py: src/openrouter/components/anthropicthinkingdisplay.py:
id: 5d63f1a15f67 id: 5d63f1a15f67
last_write_checksum: sha1:388d60bbe78d3b848744564f91b31142d7bfc7e3 last_write_checksum: sha1:388d60bbe78d3b848744564f91b31142d7bfc7e3
@@ -7102,8 +7102,8 @@ trackedFiles:
pristine_git_object: b937a2df121366183d40d9f124c9f1bf21b798db pristine_git_object: b937a2df121366183d40d9f124c9f1bf21b798db
src/openrouter/components/chatcontentcachecontrol.py: src/openrouter/components/chatcontentcachecontrol.py:
id: fa34483583b7 id: fa34483583b7
last_write_checksum: sha1:e34eec1c61fdbf41456cb25b3704f4f26c7590ac last_write_checksum: sha1:b8a998d3dc83c8edd1a35c25846f59d01ff2649c
pristine_git_object: e23d40503368f1c30d2bd3335b9196fea2094c51 pristine_git_object: d60c7b13de35893537d9c6867c1e51a888dcc72d
src/openrouter/components/chatcontentfile.py: src/openrouter/components/chatcontentfile.py:
id: 2a747b8033b8 id: 2a747b8033b8
last_write_checksum: sha1:1203efad65c0e72e18fa600bd8ebbc72a82340fa last_write_checksum: sha1:1203efad65c0e72e18fa600bd8ebbc72a82340fa
@@ -7118,8 +7118,8 @@ trackedFiles:
pristine_git_object: 61415b0c78602a7fabc4bf7f62ac869bacf38195 pristine_git_object: 61415b0c78602a7fabc4bf7f62ac869bacf38195
src/openrouter/components/chatcontenttext.py: src/openrouter/components/chatcontenttext.py:
id: 5301d8af31e7 id: 5301d8af31e7
last_write_checksum: sha1:8c30921c5e1530becb0f646866787aa3d1deab13 last_write_checksum: sha1:f4abea5885878b3c6aaacac28829fae85a7c25ce
pristine_git_object: 86d0a046b85b9caba7f628344d55febb71936ca9 pristine_git_object: 45f9d3e9873702cbffb4cd7078fff4fffe2b81d2
src/openrouter/components/chatcontentvideo.py: src/openrouter/components/chatcontentvideo.py:
id: 42859cb5a9f4 id: 42859cb5a9f4
last_write_checksum: sha1:f60354c6602613f3f8628ca29330a6037fcf20e9 last_write_checksum: sha1:f60354c6602613f3f8628ca29330a6037fcf20e9
@@ -7162,8 +7162,8 @@ trackedFiles:
pristine_git_object: 7a6140f2b8afb0cde9b031322a4284c9cbce7e33 pristine_git_object: 7a6140f2b8afb0cde9b031322a4284c9cbce7e33
src/openrouter/components/chatfunctiontool.py: src/openrouter/components/chatfunctiontool.py:
id: 1fcf92f8a7da id: 1fcf92f8a7da
last_write_checksum: sha1:a18e2ed4c62125b690f3baa254a8c1763f647726 last_write_checksum: sha1:bd0f773ec2135cbfdffb56db092587410c7e5daf
pristine_git_object: a225159f24b91219a9f3d7292feefabd81515cf6 pristine_git_object: 69f8f1c4be7afa41b812c012d3cf1b0906efcf1b
src/openrouter/components/chatjsonschemaconfig.py: src/openrouter/components/chatjsonschemaconfig.py:
id: d2278e414fab id: d2278e414fab
last_write_checksum: sha1:73791c8aa5f3dec4a25903dcc9c61118fc88a4f3 last_write_checksum: sha1:73791c8aa5f3dec4a25903dcc9c61118fc88a4f3
@@ -7182,8 +7182,8 @@ trackedFiles:
pristine_git_object: 8bbde153b53887e426989fd5b0486b6dc1122810 pristine_git_object: 8bbde153b53887e426989fd5b0486b6dc1122810
src/openrouter/components/chatrequest.py: src/openrouter/components/chatrequest.py:
id: 5e39eaefa9cf id: 5e39eaefa9cf
last_write_checksum: sha1:e6f5dca8d7972b45dbb40d94867331ad0d1e09e7 last_write_checksum: sha1:2b4baaf4c3bf6d32c3514edfaab04ded9b71b001
pristine_git_object: 914012755c34634740a1c91aa3f3f7b3da3f095e pristine_git_object: 4c79a21cfcf76239f27b251c35a5600839703c3a
src/openrouter/components/chatresult.py: src/openrouter/components/chatresult.py:
id: 9062fe2935fe id: 9062fe2935fe
last_write_checksum: sha1:2ff1963c54e7e79500115c8549c8d395f4e336d1 last_write_checksum: sha1:2ff1963c54e7e79500115c8549c8d395f4e336d1
@@ -7610,8 +7610,8 @@ trackedFiles:
pristine_git_object: 0015e263bbb2d817af89780610cc8f25ff852e73 pristine_git_object: 0015e263bbb2d817af89780610cc8f25ff852e73
src/openrouter/components/fusionservertoolconfig.py: src/openrouter/components/fusionservertoolconfig.py:
id: 0d1d229c531f id: 0d1d229c531f
last_write_checksum: sha1:df3f862acbedb4f4b3ed8b3e35f04fdf804b943f last_write_checksum: sha1:ee754be65e6d2a35ed3f1f3985a0c3600116a6d0
pristine_git_object: e89fc6db28423e964f76e352ec20aa78763a8649 pristine_git_object: f7377d2de6c4e6129e8c7f9916200a06ae862afa
src/openrouter/components/fusionsource.py: src/openrouter/components/fusionsource.py:
id: ce20a4a9f32f id: ce20a4a9f32f
last_write_checksum: sha1:2141dd1dd9396035bf8070878c9ec945331514c8 last_write_checksum: sha1:2141dd1dd9396035bf8070878c9ec945331514c8
@@ -7802,8 +7802,8 @@ trackedFiles:
pristine_git_object: 39681f4c16af7831a14dcccf7447e742c8a2b43b pristine_git_object: 39681f4c16af7831a14dcccf7447e742c8a2b43b
src/openrouter/components/inputtext.py: src/openrouter/components/inputtext.py:
id: a995ae96be99 id: a995ae96be99
last_write_checksum: sha1:f78a6a35792b704f18ee6e1312c24f78f38cca36 last_write_checksum: sha1:2da9ff645acc96772701908830151c5a707ed021
pristine_git_object: ab1dcb92e9370c9252ac5623a447669cca01e0ee pristine_git_object: 81a4151b3b1cdb6eb262a9ede28217b6c0afd1df
src/openrouter/components/inputvideo.py: src/openrouter/components/inputvideo.py:
id: d262affd6f7e id: d262affd6f7e
last_write_checksum: sha1:60c1792ab1f440ae67b5b451fd4d195e02783161 last_write_checksum: sha1:60c1792ab1f440ae67b5b451fd4d195e02783161
@@ -7922,16 +7922,16 @@ trackedFiles:
pristine_git_object: 05b041967edf0ab2d8458f9f0fc3577f0a55fcfc pristine_git_object: 05b041967edf0ab2d8458f9f0fc3577f0a55fcfc
src/openrouter/components/messagesmessageparam.py: src/openrouter/components/messagesmessageparam.py:
id: ed8bbab6720d id: ed8bbab6720d
last_write_checksum: sha1:4e5fb461702f78194596d7de570c7da25f05249e last_write_checksum: sha1:479cfffaf1013b89e92bb78ee7d5c5ad6646a1ca
pristine_git_object: 8a9738b371fb0ed0cd30d0bb7cc237fc7168eb92 pristine_git_object: e556cb06f6cf993616dd14191c43d6eca82edd7d
src/openrouter/components/messagesoutputconfig.py: src/openrouter/components/messagesoutputconfig.py:
id: 54d1e297f740 id: 54d1e297f740
last_write_checksum: sha1:01e41c73a07af6167f6934394f6e1f419f7ab878 last_write_checksum: sha1:01e41c73a07af6167f6934394f6e1f419f7ab878
pristine_git_object: 9b033dd8570e1d71ddb61a63a4b320d3003bd3ef pristine_git_object: 9b033dd8570e1d71ddb61a63a4b320d3003bd3ef
src/openrouter/components/messagesrequest.py: src/openrouter/components/messagesrequest.py:
id: 72b1f339024d id: 72b1f339024d
last_write_checksum: sha1:a14c123f9362720670a6cf4eb4bd55478d5fa376 last_write_checksum: sha1:31b9b9af3f04cb27796a56329b3ad8629c8ae33a
pristine_git_object: d2e30398f665dc6b5e8dac7f53bffab6e490d14b pristine_git_object: b4d8f657ed28a53d7cdd1dd963694631070d1e40
src/openrouter/components/messagessearchmodelsservertool.py: src/openrouter/components/messagessearchmodelsservertool.py:
id: 53d755edb120 id: 53d755edb120
last_write_checksum: sha1:389786fc22ba4bb47b0a086b7aa321f2f04fe943 last_write_checksum: sha1:389786fc22ba4bb47b0a086b7aa321f2f04fe943
@@ -8362,8 +8362,8 @@ trackedFiles:
pristine_git_object: a7c4599ac6ab794f54968f12cd6b15c3137a5159 pristine_git_object: a7c4599ac6ab794f54968f12cd6b15c3137a5159
src/openrouter/components/promptcachebreakpoint.py: src/openrouter/components/promptcachebreakpoint.py:
id: f1e589a13300 id: f1e589a13300
last_write_checksum: sha1:458ef5170c7c9ea1367dcfb090b1b28333aff73d last_write_checksum: sha1:ea0f5d81093058c261edd1937b4c7148989a203a
pristine_git_object: a082e2d9b4d27d0cffb7794279ab5fb9d9e06789 pristine_git_object: 13fbce26b5254ca66ebecfbbc8701d2002483399
src/openrouter/components/promptcacheoptions.py: src/openrouter/components/promptcacheoptions.py:
id: 5c39fe555e03 id: 5c39fe555e03
last_write_checksum: sha1:77647b94fd80f6cfcae368bff4c87cbaf903a98e last_write_checksum: sha1:77647b94fd80f6cfcae368bff4c87cbaf903a98e
@@ -8538,8 +8538,8 @@ trackedFiles:
pristine_git_object: f38c8edcb616d9e7cbd0f724fbb9d9f5aae014f9 pristine_git_object: f38c8edcb616d9e7cbd0f724fbb9d9f5aae014f9
src/openrouter/components/responsesrequest.py: src/openrouter/components/responsesrequest.py:
id: 8c850080ec5d id: 8c850080ec5d
last_write_checksum: sha1:f9f52d25db9ee9e72fbdf8491c47a3fc3a9971c7 last_write_checksum: sha1:52b22af092ebe8d61503b02aaaa806737ecd295f
pristine_git_object: 6f97e56c468a1df21fc11fd4c3310fa648968c5f pristine_git_object: 92873ed0edafcd63a858b859da883c1431f15e56
src/openrouter/components/responsesstreamingresponse.py: src/openrouter/components/responsesstreamingresponse.py:
id: 142379f3bd90 id: 142379f3bd90
last_write_checksum: sha1:a48d6f3610f5d15248ef4d25a0d88a0a4a1a9db7 last_write_checksum: sha1:a48d6f3610f5d15248ef4d25a0d88a0a4a1a9db7
@@ -9474,8 +9474,8 @@ trackedFiles:
pristine_git_object: 31c74edd544c4ab7d4c5153a6d2888623ca66aed pristine_git_object: 31c74edd544c4ab7d4c5153a6d2888623ca66aed
src/openrouter/presets.py: src/openrouter/presets.py:
id: bd0c40379dcd id: bd0c40379dcd
last_write_checksum: sha1:6de7b3adbfe270bf5b492b1da06230b9f6cc35d5 last_write_checksum: sha1:3b194964b25c57db796921e561898ae7073b6cbf
pristine_git_object: 250c20d191dfd4a999ab461d6d5d1f066e1cfbeb pristine_git_object: 832689477bac35ba16b604b330fdb3720e90a3a4
src/openrouter/providers.py: src/openrouter/providers.py:
id: debc4c48f149 id: debc4c48f149
last_write_checksum: sha1:11dbc607825edee8dcc84b39ee8171e097c47acf last_write_checksum: sha1:11dbc607825edee8dcc84b39ee8171e097c47acf
@@ -9490,8 +9490,8 @@ trackedFiles:
pristine_git_object: 76378ffba1806e0ca66fc873db1a568be0becebb pristine_git_object: 76378ffba1806e0ca66fc873db1a568be0becebb
src/openrouter/responses.py: src/openrouter/responses.py:
id: f2108fb635e1 id: f2108fb635e1
last_write_checksum: sha1:b0e886f229ea3a440be77c9a037de5ffcf824100 last_write_checksum: sha1:cea59e9d4f9ff5a5c61bc69ab696f6a82d77c137
pristine_git_object: fbf686bae92410678880d1bbdb8da82c04b1869d pristine_git_object: 74419ce01c4c30820bb1e6103d2542b59d062d7a
src/openrouter/sdk.py: src/openrouter/sdk.py:
id: ee9846c4c9c5 id: ee9846c4c9c5
last_write_checksum: sha1:2247537eca7f88a8f36e53fe011904b98569ad52 last_write_checksum: sha1:2247537eca7f88a8f36e53fe011904b98569ad52
@@ -11149,4 +11149,4 @@ examples:
"500": "500":
application/json: {"error": {"code": 500, "message": "Internal Server Error"}} application/json: {"error": {"code": 500, "message": "Internal Server Error"}}
examplesVersion: 1.0.2 examplesVersion: 1.0.2
releaseNotes: "## Python SDK Changes:\n* `open_router.guardrails.list()`: `response.data[].enforce_zdr_xai` **Added**\n* `open_router.guardrails.create()`: \n * `request.enforce_zdr_xai` **Added**\n * `response.data.enforce_zdr_xai` **Added**\n* `open_router.guardrails.get()`: `response.data.enforce_zdr_xai` **Added**\n* `open_router.guardrails.update()`: \n * `request.enforce_zdr_xai` **Added**\n * `response.data.enforce_zdr_xai` **Added**\n" releaseNotes: "## Python SDK Changes:\n* `open_router.chat.send()`: \n * `request` **Changed**\n * `response.choices[].message.content.union(Array<ChatContentItems>)[].union(text).prompt_cache_breakpoint` **Added**\n* `open_router.presets.create_presets_chat_completions()`: `request` **Changed**\n"
.speakeasy/gen.yaml
+1 -1
View File
@@ -36,7 +36,7 @@ generation:
documentation: mintlify documentation: mintlify
preApplyUnionDiscriminators: true preApplyUnionDiscriminators: true
python: python:
version: 0.11.31 version: 0.11.32
additionalDependencies: additionalDependencies:
dev: {} dev: {}
main: {} main: {}
.speakeasy/out.openapi.yaml
+11 -3
View File
@@ -671,7 +671,7 @@ components:
type: object type: object
AnthropicCacheControlDirective: AnthropicCacheControlDirective:
description: >- description: >-
Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
example: example:
type: ephemeral type: ephemeral
properties: properties:
@@ -4458,7 +4458,8 @@ components:
- $ref: '#/components/schemas/AnthropicCacheControlDirective' - $ref: '#/components/schemas/AnthropicCacheControlDirective'
- properties: {} - properties: {}
type: object type: object
description: Cache control for the content part description: >-
Anthropic-style cache breakpoint for the content part. Interchangeable with the OpenAI-style `prompt_cache_breakpoint` marker: OpenRouter converts between the two based on the provider serving the request.
example: example:
ttl: 5m ttl: 5m
type: ephemeral type: ephemeral
@@ -4553,6 +4554,8 @@ components:
properties: properties:
cache_control: cache_control:
$ref: '#/components/schemas/ChatContentCacheControl' $ref: '#/components/schemas/ChatContentCacheControl'
prompt_cache_breakpoint:
$ref: '#/components/schemas/PromptCacheBreakpoint'
text: text:
type: string type: string
type: type:
@@ -5047,6 +5050,11 @@ components:
format: double format: double
nullable: true nullable: true
type: number type: number
prompt_cache_key:
nullable: true
type: string
prompt_cache_options:
$ref: '#/components/schemas/PromptCacheOptions'
provider: provider:
$ref: '#/components/schemas/ProviderPreferences' $ref: '#/components/schemas/ProviderPreferences'
reasoning: reasoning:
@@ -18337,7 +18345,7 @@ components:
type: object type: object
PromptCacheBreakpoint: PromptCacheBreakpoint:
description: >- description: >-
Marks an explicit prompt-cache boundary on this content block. Everything through the block carrying this marker is part of the candidate cached prefix. Only supported by OpenAI GPT-5.6 and newer. Marks an explicit prompt-cache boundary on this content block (OpenAI-style). Everything through the block carrying this marker is part of the candidate cached prefix. Supported natively by OpenAI GPT-5.6 and newer; on providers that use Anthropic-style `cache_control`, OpenRouter converts the marker to that format automatically.
example: example:
mode: explicit mode: explicit
nullable: true nullable: true
.speakeasy/workflow.lock
+5 -5
View File
@@ -2,8 +2,8 @@ speakeasyVersion: 1.787.0
sources: sources:
OpenRouter API: OpenRouter API:
sourceNamespace: open-router-chat-completions-api sourceNamespace: open-router-chat-completions-api
sourceRevisionDigest: sha256:e41d9374b62c6364b29b039f24a6fa33b6e1aecd74f3d592b4e5dc0e8915f539 sourceRevisionDigest: sha256:7ba0690911aa50c921dd9903da84f01fc6f17c079b207e23cdf58ea361564199
sourceBlobDigest: sha256:3395078b604c8be9e707ef5d4bfc1861adb8b27ec5b3e960694744eb75dce17c sourceBlobDigest: sha256:d60cb272a488a05daf5bfe8efc1b69083f75852bff82d8deea1714b429b4d4fc
tags: tags:
- latest - latest
- 1.0.0 - 1.0.0
@@ -11,10 +11,10 @@ targets:
open-router: open-router:
source: OpenRouter API source: OpenRouter API
sourceNamespace: open-router-chat-completions-api sourceNamespace: open-router-chat-completions-api
sourceRevisionDigest: sha256:e41d9374b62c6364b29b039f24a6fa33b6e1aecd74f3d592b4e5dc0e8915f539 sourceRevisionDigest: sha256:7ba0690911aa50c921dd9903da84f01fc6f17c079b207e23cdf58ea361564199
sourceBlobDigest: sha256:3395078b604c8be9e707ef5d4bfc1861adb8b27ec5b3e960694744eb75dce17c sourceBlobDigest: sha256:d60cb272a488a05daf5bfe8efc1b69083f75852bff82d8deea1714b429b4d4fc
codeSamplesNamespace: open-router-python-code-samples codeSamplesNamespace: open-router-python-code-samples
codeSamplesRevisionDigest: sha256:4cb898e1201b335e38601e54fb5999ae9b324e5a45e518a0b443bd823affee26 codeSamplesRevisionDigest: sha256:f52550d8391f3b538a0b03a1f483c981fcb926e95bec37ffa581344cfaf0d1e8
workflow: workflow:
workflowVersion: 1.0.0 workflowVersion: 1.0.0
speakeasyVersion: 1.787.0 speakeasyVersion: 1.787.0
RELEASES.md
+11 -1
View File
@@ -418,4 +418,14 @@ Based on:
### Generated ### Generated
- [python v0.11.31] . - [python v0.11.31] .
### Releases ### Releases
- [PyPI v0.11.31] https://pypi.org/project/openrouter/0.11.31 - . - [PyPI v0.11.31] https://pypi.org/project/openrouter/0.11.31 - .
## 2026-07-14 21:26:36
### Changes
Based on:
- OpenAPI Doc
- Speakeasy CLI 1.787.0 (2.914.0) https://github.com/speakeasy-api/speakeasy
### Generated
- [python v0.11.32] .
### Releases
- [PyPI v0.11.32] https://pypi.org/project/openrouter/0.11.32 - .
docs/components/anthropiccachecontroldirective.mdx
+1 -1
View File
@@ -2,7 +2,7 @@
title: "AnthropicCacheControlDirective" title: "AnthropicCacheControlDirective"
--- ---
Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
## Fields ## Fields
docs/components/anthropicdocumentblockparam.mdx
+8 -8
View File
@@ -4,11 +4,11 @@ title: "AnthropicDocumentBlockParam"
## Fields ## Fields
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `citations` | [OptionalNullable[components.AnthropicDocumentBlockParamCitations]](../components/anthropicdocumentblockparamcitations.mdx) | :heavy_minus_sign: | N/A | | | `citations` | [OptionalNullable[components.AnthropicDocumentBlockParamCitations]](../components/anthropicdocumentblockparamcitations.mdx) | :heavy_minus_sign: | N/A | |
| `context` | *OptionalNullable[str]* | :heavy_minus_sign: | N/A | | | `context` | *OptionalNullable[str]* | :heavy_minus_sign: | N/A | |
| `source` | [components.AnthropicDocumentBlockParamSourceUnion](../components/anthropicdocumentblockparamsourceunion.mdx) | :heavy_check_mark: | N/A | | | `source` | [components.AnthropicDocumentBlockParamSourceUnion](../components/anthropicdocumentblockparamsourceunion.mdx) | :heavy_check_mark: | N/A | |
| `title` | *OptionalNullable[str]* | :heavy_minus_sign: | N/A | | | `title` | *OptionalNullable[str]* | :heavy_minus_sign: | N/A | |
| `type` | [components.TypeDocument](../components/typedocument.mdx) | :heavy_check_mark: | N/A | | | `type` | [components.TypeDocument](../components/typedocument.mdx) | :heavy_check_mark: | N/A | |
docs/components/anthropicimageblockparam.mdx
+5 -5
View File
@@ -4,8 +4,8 @@ title: "AnthropicImageBlockParam"
## Fields ## Fields
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `source` | [components.AnthropicImageBlockParamSource](../components/anthropicimageblockparamsource.mdx) | :heavy_check_mark: | N/A | | | `source` | [components.AnthropicImageBlockParamSource](../components/anthropicimageblockparamsource.mdx) | :heavy_check_mark: | N/A | |
| `type` | [components.AnthropicImageBlockParamType](../components/anthropicimageblockparamtype.mdx) | :heavy_check_mark: | N/A | | | `type` | [components.AnthropicImageBlockParamType](../components/anthropicimageblockparamtype.mdx) | :heavy_check_mark: | N/A | |
docs/components/anthropicsearchresultblockparam.mdx
+8 -8
View File
@@ -4,11 +4,11 @@ title: "AnthropicSearchResultBlockParam"
## Fields ## Fields
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `citations` | [Optional[components.AnthropicSearchResultBlockParamCitations]](../components/anthropicsearchresultblockparamcitations.mdx) | :heavy_minus_sign: | N/A | | | `citations` | [Optional[components.AnthropicSearchResultBlockParamCitations]](../components/anthropicsearchresultblockparamcitations.mdx) | :heavy_minus_sign: | N/A | |
| `content` | List[[components.AnthropicTextBlockParam](../components/anthropictextblockparam.mdx)] | :heavy_check_mark: | N/A | | | `content` | List[[components.AnthropicTextBlockParam](../components/anthropictextblockparam.mdx)] | :heavy_check_mark: | N/A | |
| `source` | *str* | :heavy_check_mark: | N/A | | | `source` | *str* | :heavy_check_mark: | N/A | |
| `title` | *str* | :heavy_check_mark: | N/A | | | `title` | *str* | :heavy_check_mark: | N/A | |
| `type` | [components.AnthropicSearchResultBlockParamType](../components/anthropicsearchresultblockparamtype.mdx) | :heavy_check_mark: | N/A | | | `type` | [components.AnthropicSearchResultBlockParamType](../components/anthropicsearchresultblockparamtype.mdx) | :heavy_check_mark: | N/A | |
docs/components/anthropictextblockparam.mdx
+6 -6
View File
@@ -4,9 +4,9 @@ title: "AnthropicTextBlockParam"
## Fields ## Fields
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `citations` | List[[components.Citation](../components/citation.mdx)] | :heavy_minus_sign: | N/A | | | `citations` | List[[components.Citation](../components/citation.mdx)] | :heavy_minus_sign: | N/A | |
| `text` | *str* | :heavy_check_mark: | N/A | | | `text` | *str* | :heavy_check_mark: | N/A | |
| `type` | [components.AnthropicTextBlockParamType](../components/anthropictextblockparamtype.mdx) | :heavy_check_mark: | N/A | | | `type` | [components.AnthropicTextBlockParamType](../components/anthropictextblockparamtype.mdx) | :heavy_check_mark: | N/A | |
docs/components/caching.mdx
+1 -1
View File
@@ -2,7 +2,7 @@
title: "Caching" title: "Caching"
--- ---
Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
## Fields ## Fields
docs/components/chatcontentcachecontrol.mdx
+1 -1
View File
@@ -2,7 +2,7 @@
title: "ChatContentCacheControl" title: "ChatContentCacheControl"
--- ---
Cache control for the content part Anthropic-style cache breakpoint for the content part. Interchangeable with the OpenAI-style `prompt_cache_breakpoint` marker: OpenRouter converts between the two based on the provider serving the request.
## Fields ## Fields
docs/components/chatcontenttext.mdx
+6 -5
View File
@@ -7,8 +7,9 @@ Text content part
## Fields ## Fields
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| ---------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------- | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `cache_control` | [Optional[components.ChatContentCacheControl]](../components/chatcontentcachecontrol.mdx) | :heavy_minus_sign: | Cache control for the content part | \{<br/>"ttl": "5m",<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.ChatContentCacheControl]](../components/chatcontentcachecontrol.mdx) | :heavy_minus_sign: | Anthropic-style cache breakpoint for the content part. Interchangeable with the OpenAI-style `prompt_cache_breakpoint` marker: OpenRouter converts between the two based on the provider serving the request. | \{<br/>"ttl": "5m",<br/>"type": "ephemeral"<br/>} |
| `text` | *str* | :heavy_check_mark: | N/A | | | `prompt_cache_breakpoint` | [OptionalNullable[components.PromptCacheBreakpoint]](../components/promptcachebreakpoint.mdx) | :heavy_minus_sign: | Marks an explicit prompt-cache boundary on this content block (OpenAI-style). Everything through the block carrying this marker is part of the candidate cached prefix. Supported natively by OpenAI GPT-5.6 and newer; on providers that use Anthropic-style `cache_control`, OpenRouter converts the marker to that format automatically. | \{<br/>"mode": "explicit"<br/>} |
| `type` | [components.ChatContentTextType](../components/chatcontenttexttype.mdx) | :heavy_check_mark: | N/A | | | `text` | *str* | :heavy_check_mark: | N/A | |
| `type` | [components.ChatContentTextType](../components/chatcontenttexttype.mdx) | :heavy_check_mark: | N/A | |
docs/components/chatfunctiontoolfunction.mdx
+1 -1
View File
@@ -6,6 +6,6 @@ title: "ChatFunctionToolFunction"
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `cache_control` | [Optional[components.ChatContentCacheControl]](../components/chatcontentcachecontrol.mdx) | :heavy_minus_sign: | Cache control for the content part | \{<br/>"ttl": "5m",<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.ChatContentCacheControl]](../components/chatcontentcachecontrol.mdx) | :heavy_minus_sign: | Anthropic-style cache breakpoint for the content part. Interchangeable with the OpenAI-style `prompt_cache_breakpoint` marker: OpenRouter converts between the two based on the provider serving the request. | \{<br/>"ttl": "5m",<br/>"type": "ephemeral"<br/>} |
| `function` | [components.ChatFunctionToolFunctionFunction](../components/chatfunctiontoolfunctionfunction.mdx) | :heavy_check_mark: | Function definition for tool calling | \{<br/>"description": "Get the current weather for a location",<br/>"name": "get_weather",<br/>"parameters": \{<br/>"properties": \{<br/>"location": \{<br/>"description": "City name",<br/>"type": "string"<br/>}<br/>},<br/>"required": [<br/>"location"<br/>],<br/>"type": "object"<br/>}<br/>} | | `function` | [components.ChatFunctionToolFunctionFunction](../components/chatfunctiontoolfunctionfunction.mdx) | :heavy_check_mark: | Function definition for tool calling | \{<br/>"description": "Get the current weather for a location",<br/>"name": "get_weather",<br/>"parameters": \{<br/>"properties": \{<br/>"location": \{<br/>"description": "City name",<br/>"type": "string"<br/>}<br/>},<br/>"required": [<br/>"location"<br/>],<br/>"type": "object"<br/>}<br/>} |
| `type` | [components.ChatFunctionToolType](../components/chatfunctiontooltype.mdx) | :heavy_check_mark: | N/A | | | `type` | [components.ChatFunctionToolType](../components/chatfunctiontooltype.mdx) | :heavy_check_mark: | N/A | |
docs/components/chatrequest.mdx
+3 -1
View File
@@ -9,7 +9,7 @@ Chat completion request parameters
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `debug` | [Optional[components.ChatDebugOptions]](../components/chatdebugoptions.mdx) | :heavy_minus_sign: | Debug options for inspecting request transformations (streaming only) | \{<br/>"echo_upstream_body": true<br/>} | | `debug` | [Optional[components.ChatDebugOptions]](../components/chatdebugoptions.mdx) | :heavy_minus_sign: | Debug options for inspecting request transformations (streaming only) | \{<br/>"echo_upstream_body": true<br/>} |
| `frequency_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | Frequency penalty (-2.0 to 2.0) | 0 | | `frequency_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | Frequency penalty (-2.0 to 2.0) | 0 |
| `image_config` | Dict[str, [components.ImageConfig](../components/imageconfig.mdx)] | :heavy_minus_sign: | Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. | \{<br/>"aspect_ratio": "16:9",<br/>"quality": "high"<br/>} | | `image_config` | Dict[str, [components.ImageConfig](../components/imageconfig.mdx)] | :heavy_minus_sign: | Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. | \{<br/>"aspect_ratio": "16:9",<br/>"quality": "high"<br/>} |
@@ -26,6 +26,8 @@ Chat completion request parameters
| `parallel_tool_calls` | *OptionalNullable[bool]* | :heavy_minus_sign: | Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response. | true | | `parallel_tool_calls` | *OptionalNullable[bool]* | :heavy_minus_sign: | Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response. | true |
| `plugins` | List[[components.ChatRequestPlugin](../components/chatrequestplugin.mdx)] | :heavy_minus_sign: | Plugins you want to enable for this request, including their settings. | | | `plugins` | List[[components.ChatRequestPlugin](../components/chatrequestplugin.mdx)] | :heavy_minus_sign: | Plugins you want to enable for this request, including their settings. | |
| `presence_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | Presence penalty (-2.0 to 2.0) | 0 | | `presence_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | Presence penalty (-2.0 to 2.0) | 0 |
| `prompt_cache_key` | *OptionalNullable[str]* | :heavy_minus_sign: | N/A | |
| `prompt_cache_options` | [OptionalNullable[components.PromptCacheOptions]](../components/promptcacheoptions.mdx) | :heavy_minus_sign: | Request-level prompt-cache controls. `mode: "explicit"` disables OpenAI-managed breakpoints so only blocks marked with `prompt_cache_breakpoint` are cached. Only supported by OpenAI GPT-5.6 and newer. | \{<br/>"mode": "explicit",<br/>"ttl": "30m"<br/>} |
| `provider` | [OptionalNullable[components.ProviderPreferences]](../components/providerpreferences.mdx) | :heavy_minus_sign: | When multiple model providers are available, optionally indicate your routing preference. | \{<br/>"allow_fallbacks": true<br/>} | | `provider` | [OptionalNullable[components.ProviderPreferences]](../components/providerpreferences.mdx) | :heavy_minus_sign: | When multiple model providers are available, optionally indicate your routing preference. | \{<br/>"allow_fallbacks": true<br/>} |
| `reasoning` | [Optional[components.ChatRequestReasoning]](../components/chatrequestreasoning.mdx) | :heavy_minus_sign: | Configuration options for reasoning models | \{<br/>"effort": "medium",<br/>"summary": "concise"<br/>} | | `reasoning` | [Optional[components.ChatRequestReasoning]](../components/chatrequestreasoning.mdx) | :heavy_minus_sign: | Configuration options for reasoning models | \{<br/>"effort": "medium",<br/>"summary": "concise"<br/>} |
| `reasoning_effort` | [OptionalNullable[components.ChatRequestReasoningEffort]](../components/chatrequestreasoningeffort.mdx) | :heavy_minus_sign: | Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ. | medium | | `reasoning_effort` | [OptionalNullable[components.ChatRequestReasoningEffort]](../components/chatrequestreasoningeffort.mdx) | :heavy_minus_sign: | Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ. | medium |
docs/components/contentcompaction.mdx
+5 -5
View File
@@ -4,8 +4,8 @@ title: "ContentCompaction"
## Fields ## Fields
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `content` | *Nullable[str]* | :heavy_check_mark: | N/A | | | `content` | *Nullable[str]* | :heavy_check_mark: | N/A | |
| `type` | [components.MessagesMessageParamTypeCompaction](../components/messagesmessageparamtypecompaction.mdx) | :heavy_check_mark: | N/A | | | `type` | [components.MessagesMessageParamTypeCompaction](../components/messagesmessageparamtypecompaction.mdx) | :heavy_check_mark: | N/A | |
docs/components/contentservertooluse.mdx
+7 -7
View File
@@ -4,10 +4,10 @@ title: "ContentServerToolUse"
## Fields ## Fields
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `id` | *str* | :heavy_check_mark: | N/A | | | `id` | *str* | :heavy_check_mark: | N/A | |
| `input` | *OptionalNullable[Any]* | :heavy_minus_sign: | N/A | | | `input` | *OptionalNullable[Any]* | :heavy_minus_sign: | N/A | |
| `name` | *str* | :heavy_check_mark: | N/A | | | `name` | *str* | :heavy_check_mark: | N/A | |
| `type` | [components.TypeServerToolUse](../components/typeservertooluse.mdx) | :heavy_check_mark: | N/A | | | `type` | [components.TypeServerToolUse](../components/typeservertooluse.mdx) | :heavy_check_mark: | N/A | |
docs/components/contenttoolresult.mdx
+7 -7
View File
@@ -4,10 +4,10 @@ title: "ContentToolResult"
## Fields ## Fields
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `content` | [Optional[components.MessagesMessageParamContentUnion2]](../components/messagesmessageparamcontentunion2.mdx) | :heavy_minus_sign: | N/A | | | `content` | [Optional[components.MessagesMessageParamContentUnion2]](../components/messagesmessageparamcontentunion2.mdx) | :heavy_minus_sign: | N/A | |
| `is_error` | *Optional[bool]* | :heavy_minus_sign: | N/A | | | `is_error` | *Optional[bool]* | :heavy_minus_sign: | N/A | |
| `tool_use_id` | *str* | :heavy_check_mark: | N/A | | | `tool_use_id` | *str* | :heavy_check_mark: | N/A | |
| `type` | [components.TypeToolResult](../components/typetoolresult.mdx) | :heavy_check_mark: | N/A | | | `type` | [components.TypeToolResult](../components/typetoolresult.mdx) | :heavy_check_mark: | N/A | |
docs/components/contenttooluse.mdx
+7 -7
View File
@@ -4,10 +4,10 @@ title: "ContentToolUse"
## Fields ## Fields
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `id` | *str* | :heavy_check_mark: | N/A | | | `id` | *str* | :heavy_check_mark: | N/A | |
| `input` | *OptionalNullable[Any]* | :heavy_minus_sign: | N/A | | | `input` | *OptionalNullable[Any]* | :heavy_minus_sign: | N/A | |
| `name` | *str* | :heavy_check_mark: | N/A | | | `name` | *str* | :heavy_check_mark: | N/A | |
| `type` | [components.TypeToolUse](../components/typetooluse.mdx) | :heavy_check_mark: | N/A | | | `type` | [components.TypeToolUse](../components/typetooluse.mdx) | :heavy_check_mark: | N/A | |
docs/components/contentwebsearchtoolresult.mdx
+6 -6
View File
@@ -4,9 +4,9 @@ title: "ContentWebSearchToolResult"
## Fields ## Fields
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `content` | [components.MessagesMessageParamContentUnion3](../components/messagesmessageparamcontentunion3.mdx) | :heavy_check_mark: | N/A | | | `content` | [components.MessagesMessageParamContentUnion3](../components/messagesmessageparamcontentunion3.mdx) | :heavy_check_mark: | N/A | |
| `tool_use_id` | *str* | :heavy_check_mark: | N/A | | | `tool_use_id` | *str* | :heavy_check_mark: | N/A | |
| `type` | [components.TypeWebSearchToolResult](../components/typewebsearchtoolresult.mdx) | :heavy_check_mark: | N/A | | | `type` | [components.TypeWebSearchToolResult](../components/typewebsearchtoolresult.mdx) | :heavy_check_mark: | N/A | |
docs/components/fusionservertoolconfig.mdx
+10 -10
View File
@@ -7,13 +7,13 @@ Configuration for the openrouter:fusion server tool.
## Fields ## Fields
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `analysis_models` | List[*str*] | :heavy_minus_sign: | Slugs of models to run in parallel as the analysis panel. Each model receives the user prompt with openrouter:web_search and openrouter:web_fetch enabled, then a judge model summarizes the collective output into structured analysis JSON. Capped at 8 models to bound cost amplification. Defaults to the Quality preset from /labs/fusion. | [<br/>"~anthropic/claude-opus-latest",<br/>"~openai/gpt-latest",<br/>"~google/gemini-pro-latest"<br/>] | | `analysis_models` | List[*str*] | :heavy_minus_sign: | Slugs of models to run in parallel as the analysis panel. Each model receives the user prompt with openrouter:web_search and openrouter:web_fetch enabled, then a judge model summarizes the collective output into structured analysis JSON. Capped at 8 models to bound cost amplification. Defaults to the Quality preset from /labs/fusion. | [<br/>"~anthropic/claude-opus-latest",<br/>"~openai/gpt-latest",<br/>"~google/gemini-pro-latest"<br/>] |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `max_completion_tokens` | *Optional[int]* | :heavy_minus_sign: | Maximum number of output tokens (including reasoning tokens) each panelist and the judge model may produce per inner call. Controls the total output budget so reasoning-heavy models like GPT-5.5 do not exhaust their token allowance before producing visible text. When omitted, panelists default to 32000 and the judge to 50000. | 16384 | | `max_completion_tokens` | *Optional[int]* | :heavy_minus_sign: | Maximum number of output tokens (including reasoning tokens) each panelist and the judge model may produce per inner call. Controls the total output budget so reasoning-heavy models like GPT-5.5 do not exhaust their token allowance before producing visible text. When omitted, panelists default to 32000 and the judge to 50000. | 16384 |
| `max_tool_calls` | *Optional[int]* | :heavy_minus_sign: | Maximum number of tool-calling steps each panelist (analysis model) and the judge model may take during their agentic web-research loop. Models with web_search/web_fetch enabled iterate until they produce a text response or hit this ceiling. Defaults to 8. Capped at 16. | 12 | | `max_tool_calls` | *Optional[int]* | :heavy_minus_sign: | Maximum number of tool-calling steps each panelist (analysis model) and the judge model may take during their agentic web-research loop. Models with web_search/web_fetch enabled iterate until they produce a text response or hit this ceiling. Defaults to 8. Capped at 16. | 12 |
| `model` | *Optional[str]* | :heavy_minus_sign: | Slug of the judge model that produces the structured analysis JSON. Defaults to the model used in the outer API request. | ~anthropic/claude-opus-latest | | `model` | *Optional[str]* | :heavy_minus_sign: | Slug of the judge model that produces the structured analysis JSON. Defaults to the model used in the outer API request. | ~anthropic/claude-opus-latest |
| `reasoning` | [Optional[components.FusionServerToolConfigReasoning]](../components/fusionservertoolconfigreasoning.mdx) | :heavy_minus_sign: | Reasoning configuration forwarded to panelist and judge inner calls. Use this to control reasoning effort and token budget for models that support extended thinking. | | | `reasoning` | [Optional[components.FusionServerToolConfigReasoning]](../components/fusionservertoolconfigreasoning.mdx) | :heavy_minus_sign: | Reasoning configuration forwarded to panelist and judge inner calls. Use this to control reasoning effort and token budget for models that support extended thinking. | |
| `temperature` | *Optional[float]* | :heavy_minus_sign: | Temperature forwarded to panelist inner calls. The judge always runs at temperature 0 regardless of this value. When omitted, the provider's default applies. | 0.7 | | `temperature` | *Optional[float]* | :heavy_minus_sign: | Temperature forwarded to panelist inner calls. The judge always runs at temperature 0 regardless of this value. When omitted, the provider's default applies. | 0.7 |
| `tools` | List[[components.FusionServerToolConfigTool](../components/fusionservertoolconfigtool.mdx)] | :heavy_minus_sign: | Server tools available to panelist and judge inner calls. Each entry uses the same `{ type, parameters? }` shorthand as the outer Chat Completions request. When omitted, defaults to `[{ type: "openrouter:web_search" }, { type: "openrouter:web_fetch" }]`. Pass an empty array to disable tools entirely (panelists answer from parametric knowledge only). | [<br/>\{<br/>"parameters": \{<br/>"excluded_domains": [<br/>"example.com"<br/>]<br/>},<br/>"type": "openrouter:web_search"<br/>},<br/>\{<br/>"type": "openrouter:web_fetch"<br/>}<br/>] | | `tools` | List[[components.FusionServerToolConfigTool](../components/fusionservertoolconfigtool.mdx)] | :heavy_minus_sign: | Server tools available to panelist and judge inner calls. Each entry uses the same `{ type, parameters? }` shorthand as the outer Chat Completions request. When omitted, defaults to `[{ type: "openrouter:web_search" }, { type: "openrouter:web_fetch" }]`. Pass an empty array to disable tools entirely (panelists answer from parametric knowledge only). | [<br/>\{<br/>"parameters": \{<br/>"excluded_domains": [<br/>"example.com"<br/>]<br/>},<br/>"type": "openrouter:web_search"<br/>},<br/>\{<br/>"type": "openrouter:web_fetch"<br/>}<br/>] |
docs/components/inputtext.mdx
+5 -5
View File
@@ -7,8 +7,8 @@ Text input content item
## Fields ## Fields
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `prompt_cache_breakpoint` | [OptionalNullable[components.PromptCacheBreakpoint]](../components/promptcachebreakpoint.mdx) | :heavy_minus_sign: | Marks an explicit prompt-cache boundary on this content block. Everything through the block carrying this marker is part of the candidate cached prefix. Only supported by OpenAI GPT-5.6 and newer. | \{<br/>"mode": "explicit"<br/>} | | `prompt_cache_breakpoint` | [OptionalNullable[components.PromptCacheBreakpoint]](../components/promptcachebreakpoint.mdx) | :heavy_minus_sign: | Marks an explicit prompt-cache boundary on this content block (OpenAI-style). Everything through the block carrying this marker is part of the candidate cached prefix. Supported natively by OpenAI GPT-5.6 and newer; on providers that use Anthropic-style `cache_control`, OpenRouter converts the marker to that format automatically. | \{<br/>"mode": "explicit"<br/>} |
| `text` | *str* | :heavy_check_mark: | N/A | | | `text` | *str* | :heavy_check_mark: | N/A | |
| `type` | [components.InputTextType](../components/inputtexttype.mdx) | :heavy_check_mark: | N/A | | | `type` | [components.InputTextType](../components/inputtexttype.mdx) | :heavy_check_mark: | N/A | |
docs/components/messagesrequest.mdx
+1 -1
View File
@@ -9,7 +9,7 @@ Request schema for Anthropic Messages API endpoint
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `context_management` | [OptionalNullable[components.ContextManagement]](../components/contextmanagement.mdx) | :heavy_minus_sign: | N/A | | | `context_management` | [OptionalNullable[components.ContextManagement]](../components/contextmanagement.mdx) | :heavy_minus_sign: | N/A | |
| `fallbacks` | List[[components.MessagesFallbackParam](../components/messagesfallbackparam.mdx)] | :heavy_minus_sign: | Fallback models to try if the primary model fails or refuses, in order. Handled by OpenRouter multi-model routing rather than Anthropic server-side fallbacks; cannot be combined with `models`. Each entry accepts only `model`. Maximum of 3 entries. | [<br/>\{<br/>"model": "claude-opus-4-8"<br/>}<br/>] | | `fallbacks` | List[[components.MessagesFallbackParam](../components/messagesfallbackparam.mdx)] | :heavy_minus_sign: | Fallback models to try if the primary model fails or refuses, in order. Handled by OpenRouter multi-model routing rather than Anthropic server-side fallbacks; cannot be combined with `models`. Each entry accepts only `model`. Maximum of 3 entries. | [<br/>\{<br/>"model": "claude-opus-4-8"<br/>}<br/>] |
| `max_tokens` | *Optional[int]* | :heavy_minus_sign: | N/A | | | `max_tokens` | *Optional[int]* | :heavy_minus_sign: | N/A | |
docs/components/promptcachebreakpoint.mdx
+1 -1
View File
@@ -2,7 +2,7 @@
title: "PromptCacheBreakpoint" title: "PromptCacheBreakpoint"
--- ---
Marks an explicit prompt-cache boundary on this content block. Everything through the block carrying this marker is part of the candidate cached prefix. Only supported by OpenAI GPT-5.6 and newer. Marks an explicit prompt-cache boundary on this content block (OpenAI-style). Everything through the block carrying this marker is part of the candidate cached prefix. Supported natively by OpenAI GPT-5.6 and newer; on providers that use Anthropic-style `cache_control`, OpenRouter converts the marker to that format automatically.
## Fields ## Fields
docs/components/responsesrequest.mdx
+1 -1
View File
@@ -10,7 +10,7 @@ Request schema for Responses endpoint
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `background` | *OptionalNullable[bool]* | :heavy_minus_sign: | N/A | | | `background` | *OptionalNullable[bool]* | :heavy_minus_sign: | N/A | |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `debug` | [Optional[components.ChatDebugOptions]](../components/chatdebugoptions.mdx) | :heavy_minus_sign: | Debug options for inspecting request transformations (streaming only) | \{<br/>"echo_upstream_body": true<br/>} | | `debug` | [Optional[components.ChatDebugOptions]](../components/chatdebugoptions.mdx) | :heavy_minus_sign: | Debug options for inspecting request transformations (streaming only) | \{<br/>"echo_upstream_body": true<br/>} |
| `frequency_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | N/A | | | `frequency_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | N/A | |
| `image_config` | Dict[str, [components.ImageConfig](../components/imageconfig.mdx)] | :heavy_minus_sign: | Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. | \{<br/>"aspect_ratio": "16:9",<br/>"quality": "high"<br/>} | | `image_config` | Dict[str, [components.ImageConfig](../components/imageconfig.mdx)] | :heavy_minus_sign: | Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. | \{<br/>"aspect_ratio": "16:9",<br/>"quality": "high"<br/>} |
docs/components/tooladvisor20260301.mdx
+10 -10
View File
@@ -4,13 +4,13 @@ title: "ToolAdvisor20260301"
## Fields ## Fields
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `allowed_callers` | List[[components.AnthropicAllowedCallers](../components/anthropicallowedcallers.mdx)] | :heavy_minus_sign: | N/A | [<br/>"direct"<br/>] | | `allowed_callers` | List[[components.AnthropicAllowedCallers](../components/anthropicallowedcallers.mdx)] | :heavy_minus_sign: | N/A | [<br/>"direct"<br/>] |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `caching` | [OptionalNullable[components.Caching]](../components/caching.mdx) | :heavy_minus_sign: | N/A | \{<br/>"type": "ephemeral"<br/>} | | `caching` | [OptionalNullable[components.Caching]](../components/caching.mdx) | :heavy_minus_sign: | N/A | \{<br/>"type": "ephemeral"<br/>} |
| `defer_loading` | *Optional[bool]* | :heavy_minus_sign: | N/A | | | `defer_loading` | *Optional[bool]* | :heavy_minus_sign: | N/A | |
| `max_uses` | *Optional[int]* | :heavy_minus_sign: | N/A | | | `max_uses` | *Optional[int]* | :heavy_minus_sign: | N/A | |
| `model` | *str* | :heavy_check_mark: | N/A | | | `model` | *str* | :heavy_check_mark: | N/A | |
| `name` | [components.NameAdvisor](../components/nameadvisor.mdx) | :heavy_check_mark: | N/A | | | `name` | [components.NameAdvisor](../components/nameadvisor.mdx) | :heavy_check_mark: | N/A | |
| `type` | [components.TypeAdvisor20260301](../components/typeadvisor20260301.mdx) | :heavy_check_mark: | N/A | | | `type` | [components.TypeAdvisor20260301](../components/typeadvisor20260301.mdx) | :heavy_check_mark: | N/A | |
docs/components/toolbash20250124.mdx
+5 -5
View File
@@ -4,8 +4,8 @@ title: "ToolBash20250124"
## Fields ## Fields
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `name` | [components.NameBash](../components/namebash.mdx) | :heavy_check_mark: | N/A | | | `name` | [components.NameBash](../components/namebash.mdx) | :heavy_check_mark: | N/A | |
| `type` | [components.TypeBash20250124](../components/typebash20250124.mdx) | :heavy_check_mark: | N/A | | | `type` | [components.TypeBash20250124](../components/typebash20250124.mdx) | :heavy_check_mark: | N/A | |
docs/components/toolcustom.mdx
+7 -7
View File
@@ -4,10 +4,10 @@ title: "ToolCustom"
## Fields ## Fields
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `description` | *Optional[str]* | :heavy_minus_sign: | N/A | | | `description` | *Optional[str]* | :heavy_minus_sign: | N/A | |
| `input_schema` | [components.InputSchema](../components/inputschema.mdx) | :heavy_check_mark: | N/A | | | `input_schema` | [components.InputSchema](../components/inputschema.mdx) | :heavy_check_mark: | N/A | |
| `name` | *str* | :heavy_check_mark: | N/A | | | `name` | *str* | :heavy_check_mark: | N/A | |
| `type` | [Optional[components.ToolTypeCustom]](../components/tooltypecustom.mdx) | :heavy_minus_sign: | N/A | | | `type` | [Optional[components.ToolTypeCustom]](../components/tooltypecustom.mdx) | :heavy_minus_sign: | N/A | |
docs/components/tooltexteditor20250124.mdx
+5 -5
View File
@@ -4,8 +4,8 @@ title: "ToolTextEditor20250124"
## Fields ## Fields
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `name` | [components.NameStrReplaceEditor](../components/namestrreplaceeditor.mdx) | :heavy_check_mark: | N/A | | | `name` | [components.NameStrReplaceEditor](../components/namestrreplaceeditor.mdx) | :heavy_check_mark: | N/A | |
| `type` | [components.TypeTextEditor20250124](../components/typetexteditor20250124.mdx) | :heavy_check_mark: | N/A | | | `type` | [components.TypeTextEditor20250124](../components/typetexteditor20250124.mdx) | :heavy_check_mark: | N/A | |
docs/components/toolwebsearch20250305.mdx
+9 -9
View File
@@ -4,12 +4,12 @@ title: "ToolWebSearch20250305"
## Fields ## Fields
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `allowed_domains` | List[*str*] | :heavy_minus_sign: | N/A | | | `allowed_domains` | List[*str*] | :heavy_minus_sign: | N/A | |
| `blocked_domains` | List[*str*] | :heavy_minus_sign: | N/A | | | `blocked_domains` | List[*str*] | :heavy_minus_sign: | N/A | |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `max_uses` | *OptionalNullable[int]* | :heavy_minus_sign: | N/A | | | `max_uses` | *OptionalNullable[int]* | :heavy_minus_sign: | N/A | |
| `name` | [components.NameWebSearch1](../components/namewebsearch1.mdx) | :heavy_check_mark: | N/A | | | `name` | [components.NameWebSearch1](../components/namewebsearch1.mdx) | :heavy_check_mark: | N/A | |
| `type` | [components.TypeWebSearch20250305](../components/typewebsearch20250305.mdx) | :heavy_check_mark: | N/A | | | `type` | [components.TypeWebSearch20250305](../components/typewebsearch20250305.mdx) | :heavy_check_mark: | N/A | |
| `user_location` | [OptionalNullable[components.AnthropicWebSearchToolUserLocation]](../components/anthropicwebsearchtooluserlocation.mdx) | :heavy_minus_sign: | N/A | \{<br/>"city": "San Francisco",<br/>"country": "US",<br/>"region": "California",<br/>"timezone": "America/Los_Angeles",<br/>"type": "approximate"<br/>} | | `user_location` | [OptionalNullable[components.AnthropicWebSearchToolUserLocation]](../components/anthropicwebsearchtooluserlocation.mdx) | :heavy_minus_sign: | N/A | \{<br/>"city": "San Francisco",<br/>"country": "US",<br/>"region": "California",<br/>"timezone": "America/Los_Angeles",<br/>"type": "approximate"<br/>} |
docs/components/toolwebsearch20260209.mdx
+10 -10
View File
@@ -4,13 +4,13 @@ title: "ToolWebSearch20260209"
## Fields ## Fields
| Field | Type | Required | Description | Example | | Field | Type | Required | Description | Example |
| ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `allowed_callers` | List[[components.AnthropicAllowedCallers](../components/anthropicallowedcallers.mdx)] | :heavy_minus_sign: | N/A | [<br/>"direct"<br/>] | | `allowed_callers` | List[[components.AnthropicAllowedCallers](../components/anthropicallowedcallers.mdx)] | :heavy_minus_sign: | N/A | [<br/>"direct"<br/>] |
| `allowed_domains` | List[*str*] | :heavy_minus_sign: | N/A | | | `allowed_domains` | List[*str*] | :heavy_minus_sign: | N/A | |
| `blocked_domains` | List[*str*] | :heavy_minus_sign: | N/A | | | `blocked_domains` | List[*str*] | :heavy_minus_sign: | N/A | |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `max_uses` | *OptionalNullable[int]* | :heavy_minus_sign: | N/A | | | `max_uses` | *OptionalNullable[int]* | :heavy_minus_sign: | N/A | |
| `name` | [components.NameWebSearch2](../components/namewebsearch2.mdx) | :heavy_check_mark: | N/A | | | `name` | [components.NameWebSearch2](../components/namewebsearch2.mdx) | :heavy_check_mark: | N/A | |
| `type` | [components.TypeWebSearch20260209](../components/typewebsearch20260209.mdx) | :heavy_check_mark: | N/A | | | `type` | [components.TypeWebSearch20260209](../components/typewebsearch20260209.mdx) | :heavy_check_mark: | N/A | |
| `user_location` | [OptionalNullable[components.AnthropicWebSearchToolUserLocation]](../components/anthropicwebsearchtooluserlocation.mdx) | :heavy_minus_sign: | N/A | \{<br/>"city": "San Francisco",<br/>"country": "US",<br/>"region": "California",<br/>"timezone": "America/Los_Angeles",<br/>"type": "approximate"<br/>} | | `user_location` | [OptionalNullable[components.AnthropicWebSearchToolUserLocation]](../components/anthropicwebsearchtooluserlocation.mdx) | :heavy_minus_sign: | N/A | \{<br/>"city": "San Francisco",<br/>"country": "US",<br/>"region": "California",<br/>"timezone": "America/Los_Angeles",<br/>"type": "approximate"<br/>} |
docs/sdks/chat/README.mdx
+3 -1
View File
@@ -84,7 +84,7 @@ with OpenRouter(
| `x_open_router_title` | *Optional[str]* | :heavy_minus_sign: | The app display name allows you to customize how your app appears in OpenRouter's dashboard.<br/> | | | `x_open_router_title` | *Optional[str]* | :heavy_minus_sign: | The app display name allows you to customize how your app appears in OpenRouter's dashboard.<br/> | |
| `x_open_router_categories` | *Optional[str]* | :heavy_minus_sign: | Comma-separated list of app categories (e.g. "cli-agent,cloud-agent"). Used for marketplace rankings.<br/> | | | `x_open_router_categories` | *Optional[str]* | :heavy_minus_sign: | Comma-separated list of app categories (e.g. "cli-agent,cloud-agent"). Used for marketplace rankings.<br/> | |
| `x_open_router_metadata` | [Optional[components.MetadataLevel]](../../components/metadatalevel.mdx) | :heavy_minus_sign: | Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. | enabled | | `x_open_router_metadata` | [Optional[components.MetadataLevel]](../../components/metadatalevel.mdx) | :heavy_minus_sign: | Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. | enabled |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `debug` | [Optional[components.ChatDebugOptions]](../../components/chatdebugoptions.mdx) | :heavy_minus_sign: | Debug options for inspecting request transformations (streaming only) | \{<br/>"echo_upstream_body": true<br/>} | | `debug` | [Optional[components.ChatDebugOptions]](../../components/chatdebugoptions.mdx) | :heavy_minus_sign: | Debug options for inspecting request transformations (streaming only) | \{<br/>"echo_upstream_body": true<br/>} |
| `frequency_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | Frequency penalty (-2.0 to 2.0) | 0 | | `frequency_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | Frequency penalty (-2.0 to 2.0) | 0 |
| `image_config` | Dict[str, [components.ImageConfig](../../components/imageconfig.mdx)] | :heavy_minus_sign: | Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. | \{<br/>"aspect_ratio": "16:9",<br/>"quality": "high"<br/>} | | `image_config` | Dict[str, [components.ImageConfig](../../components/imageconfig.mdx)] | :heavy_minus_sign: | Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. | \{<br/>"aspect_ratio": "16:9",<br/>"quality": "high"<br/>} |
@@ -100,6 +100,8 @@ with OpenRouter(
| `parallel_tool_calls` | *OptionalNullable[bool]* | :heavy_minus_sign: | Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response. | true | | `parallel_tool_calls` | *OptionalNullable[bool]* | :heavy_minus_sign: | Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response. | true |
| `plugins` | List[[components.ChatRequestPlugin](../../components/chatrequestplugin.mdx)] | :heavy_minus_sign: | Plugins you want to enable for this request, including their settings. | | | `plugins` | List[[components.ChatRequestPlugin](../../components/chatrequestplugin.mdx)] | :heavy_minus_sign: | Plugins you want to enable for this request, including their settings. | |
| `presence_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | Presence penalty (-2.0 to 2.0) | 0 | | `presence_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | Presence penalty (-2.0 to 2.0) | 0 |
| `prompt_cache_key` | *OptionalNullable[str]* | :heavy_minus_sign: | N/A | |
| `prompt_cache_options` | [OptionalNullable[components.PromptCacheOptions]](../../components/promptcacheoptions.mdx) | :heavy_minus_sign: | Request-level prompt-cache controls. `mode: "explicit"` disables OpenAI-managed breakpoints so only blocks marked with `prompt_cache_breakpoint` are cached. Only supported by OpenAI GPT-5.6 and newer. | \{<br/>"mode": "explicit",<br/>"ttl": "30m"<br/>} |
| `provider` | [OptionalNullable[components.ProviderPreferences]](../../components/providerpreferences.mdx) | :heavy_minus_sign: | When multiple model providers are available, optionally indicate your routing preference. | \{<br/>"allow_fallbacks": true<br/>} | | `provider` | [OptionalNullable[components.ProviderPreferences]](../../components/providerpreferences.mdx) | :heavy_minus_sign: | When multiple model providers are available, optionally indicate your routing preference. | \{<br/>"allow_fallbacks": true<br/>} |
| `reasoning` | [Optional[components.ChatRequestReasoning]](../../components/chatrequestreasoning.mdx) | :heavy_minus_sign: | Configuration options for reasoning models | \{<br/>"effort": "medium",<br/>"summary": "concise"<br/>} | | `reasoning` | [Optional[components.ChatRequestReasoning]](../../components/chatrequestreasoning.mdx) | :heavy_minus_sign: | Configuration options for reasoning models | \{<br/>"effort": "medium",<br/>"summary": "concise"<br/>} |
| `reasoning_effort` | [OptionalNullable[components.ChatRequestReasoningEffort]](../../components/chatrequestreasoningeffort.mdx) | :heavy_minus_sign: | Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ. | medium | | `reasoning_effort` | [OptionalNullable[components.ChatRequestReasoningEffort]](../../components/chatrequestreasoningeffort.mdx) | :heavy_minus_sign: | Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ. | medium |
docs/sdks/presets/README.mdx
+5 -3
View File
@@ -160,7 +160,7 @@ with OpenRouter(
| `http_referer` | *Optional[str]* | :heavy_minus_sign: | The app identifier should be your app's URL and is used as the primary identifier for rankings.<br/>This is used to track API usage per application.<br/> | | | `http_referer` | *Optional[str]* | :heavy_minus_sign: | The app identifier should be your app's URL and is used as the primary identifier for rankings.<br/>This is used to track API usage per application.<br/> | |
| `x_open_router_title` | *Optional[str]* | :heavy_minus_sign: | The app display name allows you to customize how your app appears in OpenRouter's dashboard.<br/> | | | `x_open_router_title` | *Optional[str]* | :heavy_minus_sign: | The app display name allows you to customize how your app appears in OpenRouter's dashboard.<br/> | |
| `x_open_router_categories` | *Optional[str]* | :heavy_minus_sign: | Comma-separated list of app categories (e.g. "cli-agent,cloud-agent"). Used for marketplace rankings.<br/> | | | `x_open_router_categories` | *Optional[str]* | :heavy_minus_sign: | Comma-separated list of app categories (e.g. "cli-agent,cloud-agent"). Used for marketplace rankings.<br/> | |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `debug` | [Optional[components.ChatDebugOptions]](../../components/chatdebugoptions.mdx) | :heavy_minus_sign: | Debug options for inspecting request transformations (streaming only) | \{<br/>"echo_upstream_body": true<br/>} | | `debug` | [Optional[components.ChatDebugOptions]](../../components/chatdebugoptions.mdx) | :heavy_minus_sign: | Debug options for inspecting request transformations (streaming only) | \{<br/>"echo_upstream_body": true<br/>} |
| `frequency_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | Frequency penalty (-2.0 to 2.0) | 0 | | `frequency_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | Frequency penalty (-2.0 to 2.0) | 0 |
| `image_config` | Dict[str, [components.ImageConfig](../../components/imageconfig.mdx)] | :heavy_minus_sign: | Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. | \{<br/>"aspect_ratio": "16:9",<br/>"quality": "high"<br/>} | | `image_config` | Dict[str, [components.ImageConfig](../../components/imageconfig.mdx)] | :heavy_minus_sign: | Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. | \{<br/>"aspect_ratio": "16:9",<br/>"quality": "high"<br/>} |
@@ -176,6 +176,8 @@ with OpenRouter(
| `parallel_tool_calls` | *OptionalNullable[bool]* | :heavy_minus_sign: | Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response. | true | | `parallel_tool_calls` | *OptionalNullable[bool]* | :heavy_minus_sign: | Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response. | true |
| `plugins` | List[[components.ChatRequestPlugin](../../components/chatrequestplugin.mdx)] | :heavy_minus_sign: | Plugins you want to enable for this request, including their settings. | | | `plugins` | List[[components.ChatRequestPlugin](../../components/chatrequestplugin.mdx)] | :heavy_minus_sign: | Plugins you want to enable for this request, including their settings. | |
| `presence_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | Presence penalty (-2.0 to 2.0) | 0 | | `presence_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | Presence penalty (-2.0 to 2.0) | 0 |
| `prompt_cache_key` | *OptionalNullable[str]* | :heavy_minus_sign: | N/A | |
| `prompt_cache_options` | [OptionalNullable[components.PromptCacheOptions]](../../components/promptcacheoptions.mdx) | :heavy_minus_sign: | Request-level prompt-cache controls. `mode: "explicit"` disables OpenAI-managed breakpoints so only blocks marked with `prompt_cache_breakpoint` are cached. Only supported by OpenAI GPT-5.6 and newer. | \{<br/>"mode": "explicit",<br/>"ttl": "30m"<br/>} |
| `provider` | [OptionalNullable[components.ProviderPreferences]](../../components/providerpreferences.mdx) | :heavy_minus_sign: | When multiple model providers are available, optionally indicate your routing preference. | \{<br/>"allow_fallbacks": true<br/>} | | `provider` | [OptionalNullable[components.ProviderPreferences]](../../components/providerpreferences.mdx) | :heavy_minus_sign: | When multiple model providers are available, optionally indicate your routing preference. | \{<br/>"allow_fallbacks": true<br/>} |
| `reasoning` | [Optional[components.ChatRequestReasoning]](../../components/chatrequestreasoning.mdx) | :heavy_minus_sign: | Configuration options for reasoning models | \{<br/>"effort": "medium",<br/>"summary": "concise"<br/>} | | `reasoning` | [Optional[components.ChatRequestReasoning]](../../components/chatrequestreasoning.mdx) | :heavy_minus_sign: | Configuration options for reasoning models | \{<br/>"effort": "medium",<br/>"summary": "concise"<br/>} |
| `reasoning_effort` | [OptionalNullable[components.ChatRequestReasoningEffort]](../../components/chatrequestreasoningeffort.mdx) | :heavy_minus_sign: | Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ. | medium | | `reasoning_effort` | [OptionalNullable[components.ChatRequestReasoningEffort]](../../components/chatrequestreasoningeffort.mdx) | :heavy_minus_sign: | Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ. | medium |
@@ -255,7 +257,7 @@ with OpenRouter(
| `http_referer` | *Optional[str]* | :heavy_minus_sign: | The app identifier should be your app's URL and is used as the primary identifier for rankings.<br/>This is used to track API usage per application.<br/> | | | `http_referer` | *Optional[str]* | :heavy_minus_sign: | The app identifier should be your app's URL and is used as the primary identifier for rankings.<br/>This is used to track API usage per application.<br/> | |
| `x_open_router_title` | *Optional[str]* | :heavy_minus_sign: | The app display name allows you to customize how your app appears in OpenRouter's dashboard.<br/> | | | `x_open_router_title` | *Optional[str]* | :heavy_minus_sign: | The app display name allows you to customize how your app appears in OpenRouter's dashboard.<br/> | |
| `x_open_router_categories` | *Optional[str]* | :heavy_minus_sign: | Comma-separated list of app categories (e.g. "cli-agent,cloud-agent"). Used for marketplace rankings.<br/> | | | `x_open_router_categories` | *Optional[str]* | :heavy_minus_sign: | Comma-separated list of app categories (e.g. "cli-agent,cloud-agent"). Used for marketplace rankings.<br/> | |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `context_management` | [OptionalNullable[components.ContextManagement]](../../components/contextmanagement.mdx) | :heavy_minus_sign: | N/A | | | `context_management` | [OptionalNullable[components.ContextManagement]](../../components/contextmanagement.mdx) | :heavy_minus_sign: | N/A | |
| `fallbacks` | List[[components.MessagesFallbackParam](../../components/messagesfallbackparam.mdx)] | :heavy_minus_sign: | Fallback models to try if the primary model fails or refuses, in order. Handled by OpenRouter multi-model routing rather than Anthropic server-side fallbacks; cannot be combined with `models`. Each entry accepts only `model`. Maximum of 3 entries. | [<br/>\{<br/>"model": "claude-opus-4-8"<br/>}<br/>] | | `fallbacks` | List[[components.MessagesFallbackParam](../../components/messagesfallbackparam.mdx)] | :heavy_minus_sign: | Fallback models to try if the primary model fails or refuses, in order. Handled by OpenRouter multi-model routing rather than Anthropic server-side fallbacks; cannot be combined with `models`. Each entry accepts only `model`. Maximum of 3 entries. | [<br/>\{<br/>"model": "claude-opus-4-8"<br/>}<br/>] |
| `max_tokens` | *Optional[int]* | :heavy_minus_sign: | N/A | | | `max_tokens` | *Optional[int]* | :heavy_minus_sign: | N/A | |
@@ -331,7 +333,7 @@ with OpenRouter(
| `x_open_router_title` | *Optional[str]* | :heavy_minus_sign: | The app display name allows you to customize how your app appears in OpenRouter's dashboard.<br/> | | | `x_open_router_title` | *Optional[str]* | :heavy_minus_sign: | The app display name allows you to customize how your app appears in OpenRouter's dashboard.<br/> | |
| `x_open_router_categories` | *Optional[str]* | :heavy_minus_sign: | Comma-separated list of app categories (e.g. "cli-agent,cloud-agent"). Used for marketplace rankings.<br/> | | | `x_open_router_categories` | *Optional[str]* | :heavy_minus_sign: | Comma-separated list of app categories (e.g. "cli-agent,cloud-agent"). Used for marketplace rankings.<br/> | |
| `background` | *OptionalNullable[bool]* | :heavy_minus_sign: | N/A | | | `background` | *OptionalNullable[bool]* | :heavy_minus_sign: | N/A | |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `debug` | [Optional[components.ChatDebugOptions]](../../components/chatdebugoptions.mdx) | :heavy_minus_sign: | Debug options for inspecting request transformations (streaming only) | \{<br/>"echo_upstream_body": true<br/>} | | `debug` | [Optional[components.ChatDebugOptions]](../../components/chatdebugoptions.mdx) | :heavy_minus_sign: | Debug options for inspecting request transformations (streaming only) | \{<br/>"echo_upstream_body": true<br/>} |
| `frequency_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | N/A | | | `frequency_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | N/A | |
| `image_config` | Dict[str, [components.ImageConfig](../../components/imageconfig.mdx)] | :heavy_minus_sign: | Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. | \{<br/>"aspect_ratio": "16:9",<br/>"quality": "high"<br/>} | | `image_config` | Dict[str, [components.ImageConfig](../../components/imageconfig.mdx)] | :heavy_minus_sign: | Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. | \{<br/>"aspect_ratio": "16:9",<br/>"quality": "high"<br/>} |
docs/sdks/responses/README.mdx
+1 -1
View File
@@ -69,7 +69,7 @@ with OpenRouter(
| `x_open_router_categories` | *Optional[str]* | :heavy_minus_sign: | Comma-separated list of app categories (e.g. "cli-agent,cloud-agent"). Used for marketplace rankings.<br/> | | | `x_open_router_categories` | *Optional[str]* | :heavy_minus_sign: | Comma-separated list of app categories (e.g. "cli-agent,cloud-agent"). Used for marketplace rankings.<br/> | |
| `x_open_router_metadata` | [Optional[components.MetadataLevel]](../../components/metadatalevel.mdx) | :heavy_minus_sign: | Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. | enabled | | `x_open_router_metadata` | [Optional[components.MetadataLevel]](../../components/metadatalevel.mdx) | :heavy_minus_sign: | Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. | enabled |
| `background` | *OptionalNullable[bool]* | :heavy_minus_sign: | N/A | | | `background` | *OptionalNullable[bool]* | :heavy_minus_sign: | N/A | |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. | \{<br/>"type": "ephemeral"<br/>} | | `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../../components/anthropiccachecontroldirective.mdx) | :heavy_minus_sign: | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | \{<br/>"type": "ephemeral"<br/>} |
| `debug` | [Optional[components.ChatDebugOptions]](../../components/chatdebugoptions.mdx) | :heavy_minus_sign: | Debug options for inspecting request transformations (streaming only) | \{<br/>"echo_upstream_body": true<br/>} | | `debug` | [Optional[components.ChatDebugOptions]](../../components/chatdebugoptions.mdx) | :heavy_minus_sign: | Debug options for inspecting request transformations (streaming only) | \{<br/>"echo_upstream_body": true<br/>} |
| `frequency_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | N/A | | | `frequency_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | N/A | |
| `image_config` | Dict[str, [components.ImageConfig](../../components/imageconfig.mdx)] | :heavy_minus_sign: | Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. | \{<br/>"aspect_ratio": "16:9",<br/>"quality": "high"<br/>} | | `image_config` | Dict[str, [components.ImageConfig](../../components/imageconfig.mdx)] | :heavy_minus_sign: | Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. | \{<br/>"aspect_ratio": "16:9",<br/>"quality": "high"<br/>} |
pyproject.toml
+1 -1
View File
@@ -1,6 +1,6 @@
[project] [project]
name = "openrouter" name = "openrouter"
version = "0.11.31" version = "0.11.32"
description = "Official Python Client SDK for OpenRouter." description = "Official Python Client SDK for OpenRouter."
authors = [{ name = "OpenRouter" },] authors = [{ name = "OpenRouter" },]
readme = "README-PYPI.md" readme = "README-PYPI.md"
src/openrouter/_version.py
+2 -2
View File
@@ -3,10 +3,10 @@
import importlib.metadata import importlib.metadata
__title__: str = "openrouter" __title__: str = "openrouter"
__version__: str = "0.11.31" __version__: str = "0.11.32"
__openapi_doc_version__: str = "1.0.0" __openapi_doc_version__: str = "1.0.0"
__gen_version__: str = "2.914.0" __gen_version__: str = "2.914.0"
__user_agent__: str = "speakeasy-sdk/python 0.11.31 2.914.0 1.0.0 openrouter" __user_agent__: str = "speakeasy-sdk/python 0.11.32 2.914.0 1.0.0 openrouter"
try: try:
if __package__ is not None: if __package__ is not None:
src/openrouter/chat.py
+66 -8
View File
@@ -65,6 +65,10 @@ class Chat(BaseSDK):
] ]
] = None, ] = None,
presence_penalty: OptionalNullable[float] = UNSET, presence_penalty: OptionalNullable[float] = UNSET,
prompt_cache_key: OptionalNullable[str] = UNSET,
prompt_cache_options: OptionalNullable[
Union[components.PromptCacheOptions, components.PromptCacheOptionsTypedDict]
] = UNSET,
provider: OptionalNullable[ provider: OptionalNullable[
Union[ Union[
components.ProviderPreferences, components.ProviderPreferencesTypedDict components.ProviderPreferences, components.ProviderPreferencesTypedDict
@@ -135,7 +139,7 @@ class Chat(BaseSDK):
:param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings. :param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings.
:param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. :param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility.
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: Frequency penalty (-2.0 to 2.0) :param frequency_penalty: Frequency penalty (-2.0 to 2.0)
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
@@ -151,6 +155,8 @@ class Chat(BaseSDK):
:param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response. :param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response.
:param plugins: Plugins you want to enable for this request, including their settings. :param plugins: Plugins you want to enable for this request, including their settings.
:param presence_penalty: Presence penalty (-2.0 to 2.0) :param presence_penalty: Presence penalty (-2.0 to 2.0)
:param prompt_cache_key:
:param prompt_cache_options: Request-level prompt-cache controls. `mode: \"explicit\"` disables OpenAI-managed breakpoints so only blocks marked with `prompt_cache_breakpoint` are cached. Only supported by OpenAI GPT-5.6 and newer.
:param provider: When multiple model providers are available, optionally indicate your routing preference. :param provider: When multiple model providers are available, optionally indicate your routing preference.
:param reasoning: Configuration options for reasoning models :param reasoning: Configuration options for reasoning models
:param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ. :param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ.
@@ -223,6 +229,10 @@ class Chat(BaseSDK):
] ]
] = None, ] = None,
presence_penalty: OptionalNullable[float] = UNSET, presence_penalty: OptionalNullable[float] = UNSET,
prompt_cache_key: OptionalNullable[str] = UNSET,
prompt_cache_options: OptionalNullable[
Union[components.PromptCacheOptions, components.PromptCacheOptionsTypedDict]
] = UNSET,
provider: OptionalNullable[ provider: OptionalNullable[
Union[ Union[
components.ProviderPreferences, components.ProviderPreferencesTypedDict components.ProviderPreferences, components.ProviderPreferencesTypedDict
@@ -293,7 +303,7 @@ class Chat(BaseSDK):
:param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings. :param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings.
:param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. :param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility.
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: Frequency penalty (-2.0 to 2.0) :param frequency_penalty: Frequency penalty (-2.0 to 2.0)
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
@@ -309,6 +319,8 @@ class Chat(BaseSDK):
:param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response. :param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response.
:param plugins: Plugins you want to enable for this request, including their settings. :param plugins: Plugins you want to enable for this request, including their settings.
:param presence_penalty: Presence penalty (-2.0 to 2.0) :param presence_penalty: Presence penalty (-2.0 to 2.0)
:param prompt_cache_key:
:param prompt_cache_options: Request-level prompt-cache controls. `mode: \"explicit\"` disables OpenAI-managed breakpoints so only blocks marked with `prompt_cache_breakpoint` are cached. Only supported by OpenAI GPT-5.6 and newer.
:param provider: When multiple model providers are available, optionally indicate your routing preference. :param provider: When multiple model providers are available, optionally indicate your routing preference.
:param reasoning: Configuration options for reasoning models :param reasoning: Configuration options for reasoning models
:param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ. :param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ.
@@ -381,6 +393,10 @@ class Chat(BaseSDK):
] ]
] = None, ] = None,
presence_penalty: OptionalNullable[float] = UNSET, presence_penalty: OptionalNullable[float] = UNSET,
prompt_cache_key: OptionalNullable[str] = UNSET,
prompt_cache_options: OptionalNullable[
Union[components.PromptCacheOptions, components.PromptCacheOptionsTypedDict]
] = UNSET,
provider: OptionalNullable[ provider: OptionalNullable[
Union[ Union[
components.ProviderPreferences, components.ProviderPreferencesTypedDict components.ProviderPreferences, components.ProviderPreferencesTypedDict
@@ -453,7 +469,7 @@ class Chat(BaseSDK):
:param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings. :param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings.
:param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. :param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility.
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: Frequency penalty (-2.0 to 2.0) :param frequency_penalty: Frequency penalty (-2.0 to 2.0)
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
@@ -469,6 +485,8 @@ class Chat(BaseSDK):
:param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response. :param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response.
:param plugins: Plugins you want to enable for this request, including their settings. :param plugins: Plugins you want to enable for this request, including their settings.
:param presence_penalty: Presence penalty (-2.0 to 2.0) :param presence_penalty: Presence penalty (-2.0 to 2.0)
:param prompt_cache_key:
:param prompt_cache_options: Request-level prompt-cache controls. `mode: \"explicit\"` disables OpenAI-managed breakpoints so only blocks marked with `prompt_cache_breakpoint` are cached. Only supported by OpenAI GPT-5.6 and newer.
:param provider: When multiple model providers are available, optionally indicate your routing preference. :param provider: When multiple model providers are available, optionally indicate your routing preference.
:param reasoning: Configuration options for reasoning models :param reasoning: Configuration options for reasoning models
:param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ. :param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ.
@@ -540,6 +558,10 @@ class Chat(BaseSDK):
] ]
] = None, ] = None,
presence_penalty: OptionalNullable[float] = UNSET, presence_penalty: OptionalNullable[float] = UNSET,
prompt_cache_key: OptionalNullable[str] = UNSET,
prompt_cache_options: OptionalNullable[
Union[components.PromptCacheOptions, components.PromptCacheOptionsTypedDict]
] = UNSET,
provider: OptionalNullable[ provider: OptionalNullable[
Union[ Union[
components.ProviderPreferences, components.ProviderPreferencesTypedDict components.ProviderPreferences, components.ProviderPreferencesTypedDict
@@ -612,7 +634,7 @@ class Chat(BaseSDK):
:param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings. :param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings.
:param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. :param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility.
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: Frequency penalty (-2.0 to 2.0) :param frequency_penalty: Frequency penalty (-2.0 to 2.0)
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
@@ -628,6 +650,8 @@ class Chat(BaseSDK):
:param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response. :param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response.
:param plugins: Plugins you want to enable for this request, including their settings. :param plugins: Plugins you want to enable for this request, including their settings.
:param presence_penalty: Presence penalty (-2.0 to 2.0) :param presence_penalty: Presence penalty (-2.0 to 2.0)
:param prompt_cache_key:
:param prompt_cache_options: Request-level prompt-cache controls. `mode: \"explicit\"` disables OpenAI-managed breakpoints so only blocks marked with `prompt_cache_breakpoint` are cached. Only supported by OpenAI GPT-5.6 and newer.
:param provider: When multiple model providers are available, optionally indicate your routing preference. :param provider: When multiple model providers are available, optionally indicate your routing preference.
:param reasoning: Configuration options for reasoning models :param reasoning: Configuration options for reasoning models
:param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ. :param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ.
@@ -701,6 +725,11 @@ class Chat(BaseSDK):
plugins, Optional[List[components.ChatRequestPlugin]] plugins, Optional[List[components.ChatRequestPlugin]]
), ),
presence_penalty=presence_penalty, presence_penalty=presence_penalty,
prompt_cache_key=prompt_cache_key,
prompt_cache_options=utils.get_pydantic_model(
prompt_cache_options,
OptionalNullable[components.PromptCacheOptions],
),
provider=utils.get_pydantic_model( provider=utils.get_pydantic_model(
provider, OptionalNullable[components.ProviderPreferences] provider, OptionalNullable[components.ProviderPreferences]
), ),
@@ -977,6 +1006,10 @@ class Chat(BaseSDK):
] ]
] = None, ] = None,
presence_penalty: OptionalNullable[float] = UNSET, presence_penalty: OptionalNullable[float] = UNSET,
prompt_cache_key: OptionalNullable[str] = UNSET,
prompt_cache_options: OptionalNullable[
Union[components.PromptCacheOptions, components.PromptCacheOptionsTypedDict]
] = UNSET,
provider: OptionalNullable[ provider: OptionalNullable[
Union[ Union[
components.ProviderPreferences, components.ProviderPreferencesTypedDict components.ProviderPreferences, components.ProviderPreferencesTypedDict
@@ -1047,7 +1080,7 @@ class Chat(BaseSDK):
:param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings. :param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings.
:param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. :param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility.
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: Frequency penalty (-2.0 to 2.0) :param frequency_penalty: Frequency penalty (-2.0 to 2.0)
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
@@ -1063,6 +1096,8 @@ class Chat(BaseSDK):
:param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response. :param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response.
:param plugins: Plugins you want to enable for this request, including their settings. :param plugins: Plugins you want to enable for this request, including their settings.
:param presence_penalty: Presence penalty (-2.0 to 2.0) :param presence_penalty: Presence penalty (-2.0 to 2.0)
:param prompt_cache_key:
:param prompt_cache_options: Request-level prompt-cache controls. `mode: \"explicit\"` disables OpenAI-managed breakpoints so only blocks marked with `prompt_cache_breakpoint` are cached. Only supported by OpenAI GPT-5.6 and newer.
:param provider: When multiple model providers are available, optionally indicate your routing preference. :param provider: When multiple model providers are available, optionally indicate your routing preference.
:param reasoning: Configuration options for reasoning models :param reasoning: Configuration options for reasoning models
:param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ. :param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ.
@@ -1135,6 +1170,10 @@ class Chat(BaseSDK):
] ]
] = None, ] = None,
presence_penalty: OptionalNullable[float] = UNSET, presence_penalty: OptionalNullable[float] = UNSET,
prompt_cache_key: OptionalNullable[str] = UNSET,
prompt_cache_options: OptionalNullable[
Union[components.PromptCacheOptions, components.PromptCacheOptionsTypedDict]
] = UNSET,
provider: OptionalNullable[ provider: OptionalNullable[
Union[ Union[
components.ProviderPreferences, components.ProviderPreferencesTypedDict components.ProviderPreferences, components.ProviderPreferencesTypedDict
@@ -1205,7 +1244,7 @@ class Chat(BaseSDK):
:param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings. :param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings.
:param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. :param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility.
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: Frequency penalty (-2.0 to 2.0) :param frequency_penalty: Frequency penalty (-2.0 to 2.0)
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
@@ -1221,6 +1260,8 @@ class Chat(BaseSDK):
:param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response. :param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response.
:param plugins: Plugins you want to enable for this request, including their settings. :param plugins: Plugins you want to enable for this request, including their settings.
:param presence_penalty: Presence penalty (-2.0 to 2.0) :param presence_penalty: Presence penalty (-2.0 to 2.0)
:param prompt_cache_key:
:param prompt_cache_options: Request-level prompt-cache controls. `mode: \"explicit\"` disables OpenAI-managed breakpoints so only blocks marked with `prompt_cache_breakpoint` are cached. Only supported by OpenAI GPT-5.6 and newer.
:param provider: When multiple model providers are available, optionally indicate your routing preference. :param provider: When multiple model providers are available, optionally indicate your routing preference.
:param reasoning: Configuration options for reasoning models :param reasoning: Configuration options for reasoning models
:param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ. :param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ.
@@ -1293,6 +1334,10 @@ class Chat(BaseSDK):
] ]
] = None, ] = None,
presence_penalty: OptionalNullable[float] = UNSET, presence_penalty: OptionalNullable[float] = UNSET,
prompt_cache_key: OptionalNullable[str] = UNSET,
prompt_cache_options: OptionalNullable[
Union[components.PromptCacheOptions, components.PromptCacheOptionsTypedDict]
] = UNSET,
provider: OptionalNullable[ provider: OptionalNullable[
Union[ Union[
components.ProviderPreferences, components.ProviderPreferencesTypedDict components.ProviderPreferences, components.ProviderPreferencesTypedDict
@@ -1366,7 +1411,7 @@ class Chat(BaseSDK):
:param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings. :param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings.
:param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. :param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility.
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: Frequency penalty (-2.0 to 2.0) :param frequency_penalty: Frequency penalty (-2.0 to 2.0)
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
@@ -1382,6 +1427,8 @@ class Chat(BaseSDK):
:param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response. :param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response.
:param plugins: Plugins you want to enable for this request, including their settings. :param plugins: Plugins you want to enable for this request, including their settings.
:param presence_penalty: Presence penalty (-2.0 to 2.0) :param presence_penalty: Presence penalty (-2.0 to 2.0)
:param prompt_cache_key:
:param prompt_cache_options: Request-level prompt-cache controls. `mode: \"explicit\"` disables OpenAI-managed breakpoints so only blocks marked with `prompt_cache_breakpoint` are cached. Only supported by OpenAI GPT-5.6 and newer.
:param provider: When multiple model providers are available, optionally indicate your routing preference. :param provider: When multiple model providers are available, optionally indicate your routing preference.
:param reasoning: Configuration options for reasoning models :param reasoning: Configuration options for reasoning models
:param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ. :param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ.
@@ -1453,6 +1500,10 @@ class Chat(BaseSDK):
] ]
] = None, ] = None,
presence_penalty: OptionalNullable[float] = UNSET, presence_penalty: OptionalNullable[float] = UNSET,
prompt_cache_key: OptionalNullable[str] = UNSET,
prompt_cache_options: OptionalNullable[
Union[components.PromptCacheOptions, components.PromptCacheOptionsTypedDict]
] = UNSET,
provider: OptionalNullable[ provider: OptionalNullable[
Union[ Union[
components.ProviderPreferences, components.ProviderPreferencesTypedDict components.ProviderPreferences, components.ProviderPreferencesTypedDict
@@ -1526,7 +1577,7 @@ class Chat(BaseSDK):
:param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings. :param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings.
:param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. :param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility.
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: Frequency penalty (-2.0 to 2.0) :param frequency_penalty: Frequency penalty (-2.0 to 2.0)
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
@@ -1542,6 +1593,8 @@ class Chat(BaseSDK):
:param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response. :param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response.
:param plugins: Plugins you want to enable for this request, including their settings. :param plugins: Plugins you want to enable for this request, including their settings.
:param presence_penalty: Presence penalty (-2.0 to 2.0) :param presence_penalty: Presence penalty (-2.0 to 2.0)
:param prompt_cache_key:
:param prompt_cache_options: Request-level prompt-cache controls. `mode: \"explicit\"` disables OpenAI-managed breakpoints so only blocks marked with `prompt_cache_breakpoint` are cached. Only supported by OpenAI GPT-5.6 and newer.
:param provider: When multiple model providers are available, optionally indicate your routing preference. :param provider: When multiple model providers are available, optionally indicate your routing preference.
:param reasoning: Configuration options for reasoning models :param reasoning: Configuration options for reasoning models
:param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ. :param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ.
@@ -1615,6 +1668,11 @@ class Chat(BaseSDK):
plugins, Optional[List[components.ChatRequestPlugin]] plugins, Optional[List[components.ChatRequestPlugin]]
), ),
presence_penalty=presence_penalty, presence_penalty=presence_penalty,
prompt_cache_key=prompt_cache_key,
prompt_cache_options=utils.get_pydantic_model(
prompt_cache_options,
OptionalNullable[components.PromptCacheOptions],
),
provider=utils.get_pydantic_model( provider=utils.get_pydantic_model(
provider, OptionalNullable[components.ProviderPreferences] provider, OptionalNullable[components.ProviderPreferences]
), ),
src/openrouter/components/anthropiccachecontroldirective.py
+2 -2
View File
@@ -12,14 +12,14 @@ AnthropicCacheControlDirectiveType = Literal["ephemeral",]
class AnthropicCacheControlDirectiveTypedDict(TypedDict): class AnthropicCacheControlDirectiveTypedDict(TypedDict):
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
type: AnthropicCacheControlDirectiveType type: AnthropicCacheControlDirectiveType
ttl: NotRequired[AnthropicCacheControlTTL] ttl: NotRequired[AnthropicCacheControlTTL]
class AnthropicCacheControlDirective(BaseModel): class AnthropicCacheControlDirective(BaseModel):
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
type: AnthropicCacheControlDirectiveType type: AnthropicCacheControlDirectiveType
src/openrouter/components/anthropicdocumentblockparam.py
+2 -2
View File
@@ -135,7 +135,7 @@ class AnthropicDocumentBlockParamTypedDict(TypedDict):
source: AnthropicDocumentBlockParamSourceUnionTypedDict source: AnthropicDocumentBlockParamSourceUnionTypedDict
type: TypeDocument type: TypeDocument
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
citations: NotRequired[Nullable[AnthropicDocumentBlockParamCitationsTypedDict]] citations: NotRequired[Nullable[AnthropicDocumentBlockParamCitationsTypedDict]]
context: NotRequired[Nullable[str]] context: NotRequired[Nullable[str]]
title: NotRequired[Nullable[str]] title: NotRequired[Nullable[str]]
@@ -147,7 +147,7 @@ class AnthropicDocumentBlockParam(BaseModel):
type: TypeDocument type: TypeDocument
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
citations: OptionalNullable[AnthropicDocumentBlockParamCitations] = UNSET citations: OptionalNullable[AnthropicDocumentBlockParamCitations] = UNSET
src/openrouter/components/anthropicimageblockparam.py
+2 -2
View File
@@ -42,7 +42,7 @@ class AnthropicImageBlockParamTypedDict(TypedDict):
source: AnthropicImageBlockParamSourceTypedDict source: AnthropicImageBlockParamSourceTypedDict
type: AnthropicImageBlockParamType type: AnthropicImageBlockParamType
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
class AnthropicImageBlockParam(BaseModel): class AnthropicImageBlockParam(BaseModel):
@@ -51,7 +51,7 @@ class AnthropicImageBlockParam(BaseModel):
type: AnthropicImageBlockParamType type: AnthropicImageBlockParamType
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
@model_serializer(mode="wrap") @model_serializer(mode="wrap")
def serialize_model(self, handler): def serialize_model(self, handler):
src/openrouter/components/anthropicsearchresultblockparam.py
+2 -2
View File
@@ -48,7 +48,7 @@ class AnthropicSearchResultBlockParamTypedDict(TypedDict):
title: str title: str
type: AnthropicSearchResultBlockParamType type: AnthropicSearchResultBlockParamType
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
citations: NotRequired[AnthropicSearchResultBlockParamCitationsTypedDict] citations: NotRequired[AnthropicSearchResultBlockParamCitationsTypedDict]
@@ -62,7 +62,7 @@ class AnthropicSearchResultBlockParam(BaseModel):
type: AnthropicSearchResultBlockParamType type: AnthropicSearchResultBlockParamType
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
citations: Optional[AnthropicSearchResultBlockParamCitations] = None citations: Optional[AnthropicSearchResultBlockParamCitations] = None
src/openrouter/components/anthropictextblockparam.py
+2 -2
View File
@@ -76,7 +76,7 @@ class AnthropicTextBlockParamTypedDict(TypedDict):
text: str text: str
type: AnthropicTextBlockParamType type: AnthropicTextBlockParamType
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
citations: NotRequired[Nullable[List[CitationTypedDict]]] citations: NotRequired[Nullable[List[CitationTypedDict]]]
@@ -86,7 +86,7 @@ class AnthropicTextBlockParam(BaseModel):
type: AnthropicTextBlockParamType type: AnthropicTextBlockParamType
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
citations: OptionalNullable[List[Citation]] = UNSET citations: OptionalNullable[List[Citation]] = UNSET
src/openrouter/components/chatcontentcachecontrol.py
+2 -2
View File
@@ -12,14 +12,14 @@ ChatContentCacheControlType = Literal["ephemeral",]
class ChatContentCacheControlTypedDict(TypedDict): class ChatContentCacheControlTypedDict(TypedDict):
r"""Cache control for the content part""" r"""Anthropic-style cache breakpoint for the content part. Interchangeable with the OpenAI-style `prompt_cache_breakpoint` marker: OpenRouter converts between the two based on the provider serving the request."""
type: ChatContentCacheControlType type: ChatContentCacheControlType
ttl: NotRequired[AnthropicCacheControlTTL] ttl: NotRequired[AnthropicCacheControlTTL]
class ChatContentCacheControl(BaseModel): class ChatContentCacheControl(BaseModel):
r"""Cache control for the content part""" r"""Anthropic-style cache breakpoint for the content part. Interchangeable with the OpenAI-style `prompt_cache_breakpoint` marker: OpenRouter converts between the two based on the provider serving the request."""
type: ChatContentCacheControlType type: ChatContentCacheControlType
src/openrouter/components/chatcontenttext.py
+26 -5
View File
@@ -5,7 +5,14 @@ from .chatcontentcachecontrol import (
ChatContentCacheControl, ChatContentCacheControl,
ChatContentCacheControlTypedDict, ChatContentCacheControlTypedDict,
) )
from openrouter.types import BaseModel, UNSET_SENTINEL from .promptcachebreakpoint import PromptCacheBreakpoint, PromptCacheBreakpointTypedDict
from openrouter.types import (
BaseModel,
Nullable,
OptionalNullable,
UNSET,
UNSET_SENTINEL,
)
from pydantic import model_serializer from pydantic import model_serializer
from typing import Literal, Optional from typing import Literal, Optional
from typing_extensions import NotRequired, TypedDict from typing_extensions import NotRequired, TypedDict
@@ -20,7 +27,9 @@ class ChatContentTextTypedDict(TypedDict):
text: str text: str
type: ChatContentTextType type: ChatContentTextType
cache_control: NotRequired[ChatContentCacheControlTypedDict] cache_control: NotRequired[ChatContentCacheControlTypedDict]
r"""Cache control for the content part""" r"""Anthropic-style cache breakpoint for the content part. Interchangeable with the OpenAI-style `prompt_cache_breakpoint` marker: OpenRouter converts between the two based on the provider serving the request."""
prompt_cache_breakpoint: NotRequired[Nullable[PromptCacheBreakpointTypedDict]]
r"""Marks an explicit prompt-cache boundary on this content block (OpenAI-style). Everything through the block carrying this marker is part of the candidate cached prefix. Supported natively by OpenAI GPT-5.6 and newer; on providers that use Anthropic-style `cache_control`, OpenRouter converts the marker to that format automatically."""
class ChatContentText(BaseModel): class ChatContentText(BaseModel):
@@ -31,20 +40,32 @@ class ChatContentText(BaseModel):
type: ChatContentTextType type: ChatContentTextType
cache_control: Optional[ChatContentCacheControl] = None cache_control: Optional[ChatContentCacheControl] = None
r"""Cache control for the content part""" r"""Anthropic-style cache breakpoint for the content part. Interchangeable with the OpenAI-style `prompt_cache_breakpoint` marker: OpenRouter converts between the two based on the provider serving the request."""
prompt_cache_breakpoint: OptionalNullable[PromptCacheBreakpoint] = UNSET
r"""Marks an explicit prompt-cache boundary on this content block (OpenAI-style). Everything through the block carrying this marker is part of the candidate cached prefix. Supported natively by OpenAI GPT-5.6 and newer; on providers that use Anthropic-style `cache_control`, OpenRouter converts the marker to that format automatically."""
@model_serializer(mode="wrap") @model_serializer(mode="wrap")
def serialize_model(self, handler): def serialize_model(self, handler):
optional_fields = set(["cache_control"]) optional_fields = set(["cache_control", "prompt_cache_breakpoint"])
nullable_fields = set(["prompt_cache_breakpoint"])
serialized = handler(self) serialized = handler(self)
m = {} m = {}
for n, f in type(self).model_fields.items(): for n, f in type(self).model_fields.items():
k = f.alias or n k = f.alias or n
val = serialized.get(k, serialized.get(n)) val = serialized.get(k, serialized.get(n))
is_nullable_and_explicitly_set = (
k in nullable_fields
and (self.__pydantic_fields_set__.intersection({n})) # pylint: disable=no-member
)
if val != UNSET_SENTINEL: if val != UNSET_SENTINEL:
if val is not None or k not in optional_fields: if (
val is not None
or k not in optional_fields
or is_nullable_and_explicitly_set
):
m[k] = val m[k] = val
return m return m
src/openrouter/components/chatfunctiontool.py
+2 -2
View File
@@ -111,7 +111,7 @@ class ChatFunctionToolFunctionTypedDict(TypedDict):
r"""Function definition for tool calling""" r"""Function definition for tool calling"""
type: ChatFunctionToolType type: ChatFunctionToolType
cache_control: NotRequired[ChatContentCacheControlTypedDict] cache_control: NotRequired[ChatContentCacheControlTypedDict]
r"""Cache control for the content part""" r"""Anthropic-style cache breakpoint for the content part. Interchangeable with the OpenAI-style `prompt_cache_breakpoint` marker: OpenRouter converts between the two based on the provider serving the request."""
class ChatFunctionToolFunction(BaseModel): class ChatFunctionToolFunction(BaseModel):
@@ -121,7 +121,7 @@ class ChatFunctionToolFunction(BaseModel):
type: ChatFunctionToolType type: ChatFunctionToolType
cache_control: Optional[ChatContentCacheControl] = None cache_control: Optional[ChatContentCacheControl] = None
r"""Cache control for the content part""" r"""Anthropic-style cache breakpoint for the content part. Interchangeable with the OpenAI-style `prompt_cache_breakpoint` marker: OpenRouter converts between the two based on the provider serving the request."""
@model_serializer(mode="wrap") @model_serializer(mode="wrap")
def serialize_model(self, handler): def serialize_model(self, handler):
src/openrouter/components/chatrequest.py
+15 -2
View File
@@ -38,6 +38,7 @@ from .fusionplugin import FusionPlugin, FusionPluginTypedDict
from .imageconfig import ImageConfig, ImageConfigTypedDict from .imageconfig import ImageConfig, ImageConfigTypedDict
from .moderationplugin import ModerationPlugin, ModerationPluginTypedDict from .moderationplugin import ModerationPlugin, ModerationPluginTypedDict
from .paretorouterplugin import ParetoRouterPlugin, ParetoRouterPluginTypedDict from .paretorouterplugin import ParetoRouterPlugin, ParetoRouterPluginTypedDict
from .promptcacheoptions import PromptCacheOptions, PromptCacheOptionsTypedDict
from .providerpreferences import ProviderPreferences, ProviderPreferencesTypedDict from .providerpreferences import ProviderPreferences, ProviderPreferencesTypedDict
from .responsehealingplugin import ResponseHealingPlugin, ResponseHealingPluginTypedDict from .responsehealingplugin import ResponseHealingPlugin, ResponseHealingPluginTypedDict
from .stopservertoolswhencondition import ( from .stopservertoolswhencondition import (
@@ -228,7 +229,7 @@ class ChatRequestTypedDict(TypedDict):
messages: List[ChatMessagesTypedDict] messages: List[ChatMessagesTypedDict]
r"""List of messages for the conversation""" r"""List of messages for the conversation"""
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
debug: NotRequired[ChatDebugOptionsTypedDict] debug: NotRequired[ChatDebugOptionsTypedDict]
r"""Debug options for inspecting request transformations (streaming only)""" r"""Debug options for inspecting request transformations (streaming only)"""
frequency_penalty: NotRequired[Nullable[float]] frequency_penalty: NotRequired[Nullable[float]]
@@ -259,6 +260,9 @@ class ChatRequestTypedDict(TypedDict):
r"""Plugins you want to enable for this request, including their settings.""" r"""Plugins you want to enable for this request, including their settings."""
presence_penalty: NotRequired[Nullable[float]] presence_penalty: NotRequired[Nullable[float]]
r"""Presence penalty (-2.0 to 2.0)""" r"""Presence penalty (-2.0 to 2.0)"""
prompt_cache_key: NotRequired[Nullable[str]]
prompt_cache_options: NotRequired[Nullable[PromptCacheOptionsTypedDict]]
r"""Request-level prompt-cache controls. `mode: \"explicit\"` disables OpenAI-managed breakpoints so only blocks marked with `prompt_cache_breakpoint` are cached. Only supported by OpenAI GPT-5.6 and newer."""
provider: NotRequired[Nullable[ProviderPreferencesTypedDict]] provider: NotRequired[Nullable[ProviderPreferencesTypedDict]]
r"""When multiple model providers are available, optionally indicate your routing preference.""" r"""When multiple model providers are available, optionally indicate your routing preference."""
reasoning: NotRequired[ChatRequestReasoningTypedDict] reasoning: NotRequired[ChatRequestReasoningTypedDict]
@@ -310,7 +314,7 @@ class ChatRequest(BaseModel):
r"""List of messages for the conversation""" r"""List of messages for the conversation"""
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
debug: Optional[ChatDebugOptions] = None debug: Optional[ChatDebugOptions] = None
r"""Debug options for inspecting request transformations (streaming only)""" r"""Debug options for inspecting request transformations (streaming only)"""
@@ -357,6 +361,11 @@ class ChatRequest(BaseModel):
presence_penalty: OptionalNullable[float] = UNSET presence_penalty: OptionalNullable[float] = UNSET
r"""Presence penalty (-2.0 to 2.0)""" r"""Presence penalty (-2.0 to 2.0)"""
prompt_cache_key: OptionalNullable[str] = UNSET
prompt_cache_options: OptionalNullable[PromptCacheOptions] = UNSET
r"""Request-level prompt-cache controls. `mode: \"explicit\"` disables OpenAI-managed breakpoints so only blocks marked with `prompt_cache_breakpoint` are cached. Only supported by OpenAI GPT-5.6 and newer."""
provider: OptionalNullable[ProviderPreferences] = UNSET provider: OptionalNullable[ProviderPreferences] = UNSET
r"""When multiple model providers are available, optionally indicate your routing preference.""" r"""When multiple model providers are available, optionally indicate your routing preference."""
@@ -440,6 +449,8 @@ class ChatRequest(BaseModel):
"parallel_tool_calls", "parallel_tool_calls",
"plugins", "plugins",
"presence_penalty", "presence_penalty",
"prompt_cache_key",
"prompt_cache_options",
"provider", "provider",
"reasoning", "reasoning",
"reasoning_effort", "reasoning_effort",
@@ -473,6 +484,8 @@ class ChatRequest(BaseModel):
"min_p", "min_p",
"parallel_tool_calls", "parallel_tool_calls",
"presence_penalty", "presence_penalty",
"prompt_cache_key",
"prompt_cache_options",
"provider", "provider",
"reasoning_effort", "reasoning_effort",
"repetition_penalty", "repetition_penalty",
src/openrouter/components/fusionservertoolconfig.py
+2 -2
View File
@@ -98,7 +98,7 @@ class FusionServerToolConfigTypedDict(TypedDict):
analysis_models: NotRequired[List[str]] analysis_models: NotRequired[List[str]]
r"""Slugs of models to run in parallel as the analysis panel. Each model receives the user prompt with openrouter:web_search and openrouter:web_fetch enabled, then a judge model summarizes the collective output into structured analysis JSON. Capped at 8 models to bound cost amplification. Defaults to the Quality preset from /labs/fusion.""" r"""Slugs of models to run in parallel as the analysis panel. Each model receives the user prompt with openrouter:web_search and openrouter:web_fetch enabled, then a judge model summarizes the collective output into structured analysis JSON. Capped at 8 models to bound cost amplification. Defaults to the Quality preset from /labs/fusion."""
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
max_completion_tokens: NotRequired[int] max_completion_tokens: NotRequired[int]
r"""Maximum number of output tokens (including reasoning tokens) each panelist and the judge model may produce per inner call. Controls the total output budget so reasoning-heavy models like GPT-5.5 do not exhaust their token allowance before producing visible text. When omitted, panelists default to 32000 and the judge to 50000.""" r"""Maximum number of output tokens (including reasoning tokens) each panelist and the judge model may produce per inner call. Controls the total output budget so reasoning-heavy models like GPT-5.5 do not exhaust their token allowance before producing visible text. When omitted, panelists default to 32000 and the judge to 50000."""
max_tool_calls: NotRequired[int] max_tool_calls: NotRequired[int]
@@ -120,7 +120,7 @@ class FusionServerToolConfig(BaseModel):
r"""Slugs of models to run in parallel as the analysis panel. Each model receives the user prompt with openrouter:web_search and openrouter:web_fetch enabled, then a judge model summarizes the collective output into structured analysis JSON. Capped at 8 models to bound cost amplification. Defaults to the Quality preset from /labs/fusion.""" r"""Slugs of models to run in parallel as the analysis panel. Each model receives the user prompt with openrouter:web_search and openrouter:web_fetch enabled, then a judge model summarizes the collective output into structured analysis JSON. Capped at 8 models to bound cost amplification. Defaults to the Quality preset from /labs/fusion."""
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
max_completion_tokens: Optional[int] = None max_completion_tokens: Optional[int] = None
r"""Maximum number of output tokens (including reasoning tokens) each panelist and the judge model may produce per inner call. Controls the total output budget so reasoning-heavy models like GPT-5.5 do not exhaust their token allowance before producing visible text. When omitted, panelists default to 32000 and the judge to 50000.""" r"""Maximum number of output tokens (including reasoning tokens) each panelist and the judge model may produce per inner call. Controls the total output budget so reasoning-heavy models like GPT-5.5 do not exhaust their token allowance before producing visible text. When omitted, panelists default to 32000 and the judge to 50000."""
src/openrouter/components/inputtext.py
+2 -2
View File
@@ -23,7 +23,7 @@ class InputTextTypedDict(TypedDict):
text: str text: str
type: InputTextType type: InputTextType
prompt_cache_breakpoint: NotRequired[Nullable[PromptCacheBreakpointTypedDict]] prompt_cache_breakpoint: NotRequired[Nullable[PromptCacheBreakpointTypedDict]]
r"""Marks an explicit prompt-cache boundary on this content block. Everything through the block carrying this marker is part of the candidate cached prefix. Only supported by OpenAI GPT-5.6 and newer.""" r"""Marks an explicit prompt-cache boundary on this content block (OpenAI-style). Everything through the block carrying this marker is part of the candidate cached prefix. Supported natively by OpenAI GPT-5.6 and newer; on providers that use Anthropic-style `cache_control`, OpenRouter converts the marker to that format automatically."""
class InputText(BaseModel): class InputText(BaseModel):
@@ -34,7 +34,7 @@ class InputText(BaseModel):
type: InputTextType type: InputTextType
prompt_cache_breakpoint: OptionalNullable[PromptCacheBreakpoint] = UNSET prompt_cache_breakpoint: OptionalNullable[PromptCacheBreakpoint] = UNSET
r"""Marks an explicit prompt-cache boundary on this content block. Everything through the block carrying this marker is part of the candidate cached prefix. Only supported by OpenAI GPT-5.6 and newer.""" r"""Marks an explicit prompt-cache boundary on this content block (OpenAI-style). Everything through the block carrying this marker is part of the candidate cached prefix. Supported natively by OpenAI GPT-5.6 and newer; on providers that use Anthropic-style `cache_control`, OpenRouter converts the marker to that format automatically."""
@model_serializer(mode="wrap") @model_serializer(mode="wrap")
def serialize_model(self, handler): def serialize_model(self, handler):
src/openrouter/components/messagesmessageparam.py
+10 -10
View File
@@ -50,7 +50,7 @@ class ContentCompactionTypedDict(TypedDict):
content: Nullable[str] content: Nullable[str]
type: MessagesMessageParamTypeCompaction type: MessagesMessageParamTypeCompaction
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
class ContentCompaction(BaseModel): class ContentCompaction(BaseModel):
@@ -59,7 +59,7 @@ class ContentCompaction(BaseModel):
type: MessagesMessageParamTypeCompaction type: MessagesMessageParamTypeCompaction
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
@model_serializer(mode="wrap") @model_serializer(mode="wrap")
def serialize_model(self, handler): def serialize_model(self, handler):
@@ -136,7 +136,7 @@ class ContentWebSearchToolResultTypedDict(TypedDict):
tool_use_id: str tool_use_id: str
type: TypeWebSearchToolResult type: TypeWebSearchToolResult
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
class ContentWebSearchToolResult(BaseModel): class ContentWebSearchToolResult(BaseModel):
@@ -147,7 +147,7 @@ class ContentWebSearchToolResult(BaseModel):
type: TypeWebSearchToolResult type: TypeWebSearchToolResult
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
@model_serializer(mode="wrap") @model_serializer(mode="wrap")
def serialize_model(self, handler): def serialize_model(self, handler):
@@ -174,7 +174,7 @@ class ContentServerToolUseTypedDict(TypedDict):
name: str name: str
type: TypeServerToolUse type: TypeServerToolUse
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
input: NotRequired[Nullable[Any]] input: NotRequired[Nullable[Any]]
@@ -186,7 +186,7 @@ class ContentServerToolUse(BaseModel):
type: TypeServerToolUse type: TypeServerToolUse
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
input: OptionalNullable[Any] = UNSET input: OptionalNullable[Any] = UNSET
@@ -304,7 +304,7 @@ class ContentToolResultTypedDict(TypedDict):
tool_use_id: str tool_use_id: str
type: TypeToolResult type: TypeToolResult
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
content: NotRequired[MessagesMessageParamContentUnion2TypedDict] content: NotRequired[MessagesMessageParamContentUnion2TypedDict]
is_error: NotRequired[bool] is_error: NotRequired[bool]
@@ -315,7 +315,7 @@ class ContentToolResult(BaseModel):
type: TypeToolResult type: TypeToolResult
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
content: Optional[MessagesMessageParamContentUnion2] = None content: Optional[MessagesMessageParamContentUnion2] = None
@@ -346,7 +346,7 @@ class ContentToolUseTypedDict(TypedDict):
name: str name: str
type: TypeToolUse type: TypeToolUse
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
input: NotRequired[Nullable[Any]] input: NotRequired[Nullable[Any]]
@@ -358,7 +358,7 @@ class ContentToolUse(BaseModel):
type: TypeToolUse type: TypeToolUse
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
input: OptionalNullable[Any] = UNSET input: OptionalNullable[Any] = UNSET
src/openrouter/components/messagesrequest.py
+16 -16
View File
@@ -659,14 +659,14 @@ ToolTypeEphemeral = Literal["ephemeral",]
class CachingTypedDict(TypedDict): class CachingTypedDict(TypedDict):
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
type: ToolTypeEphemeral type: ToolTypeEphemeral
ttl: NotRequired[AnthropicCacheControlTTL] ttl: NotRequired[AnthropicCacheControlTTL]
class Caching(BaseModel): class Caching(BaseModel):
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
type: ToolTypeEphemeral type: ToolTypeEphemeral
@@ -701,7 +701,7 @@ class ToolAdvisor20260301TypedDict(TypedDict):
type: TypeAdvisor20260301 type: TypeAdvisor20260301
allowed_callers: NotRequired[List[AnthropicAllowedCallers]] allowed_callers: NotRequired[List[AnthropicAllowedCallers]]
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
caching: NotRequired[Nullable[CachingTypedDict]] caching: NotRequired[Nullable[CachingTypedDict]]
defer_loading: NotRequired[bool] defer_loading: NotRequired[bool]
max_uses: NotRequired[int] max_uses: NotRequired[int]
@@ -717,7 +717,7 @@ class ToolAdvisor20260301(BaseModel):
allowed_callers: Optional[List[AnthropicAllowedCallers]] = None allowed_callers: Optional[List[AnthropicAllowedCallers]] = None
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
caching: OptionalNullable[Caching] = UNSET caching: OptionalNullable[Caching] = UNSET
@@ -766,7 +766,7 @@ class ToolWebSearch20260209TypedDict(TypedDict):
allowed_domains: NotRequired[Nullable[List[str]]] allowed_domains: NotRequired[Nullable[List[str]]]
blocked_domains: NotRequired[Nullable[List[str]]] blocked_domains: NotRequired[Nullable[List[str]]]
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
max_uses: NotRequired[Nullable[int]] max_uses: NotRequired[Nullable[int]]
user_location: NotRequired[Nullable[AnthropicWebSearchToolUserLocationTypedDict]] user_location: NotRequired[Nullable[AnthropicWebSearchToolUserLocationTypedDict]]
@@ -783,7 +783,7 @@ class ToolWebSearch20260209(BaseModel):
blocked_domains: OptionalNullable[List[str]] = UNSET blocked_domains: OptionalNullable[List[str]] = UNSET
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
max_uses: OptionalNullable[int] = UNSET max_uses: OptionalNullable[int] = UNSET
@@ -838,7 +838,7 @@ class ToolWebSearch20250305TypedDict(TypedDict):
allowed_domains: NotRequired[Nullable[List[str]]] allowed_domains: NotRequired[Nullable[List[str]]]
blocked_domains: NotRequired[Nullable[List[str]]] blocked_domains: NotRequired[Nullable[List[str]]]
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
max_uses: NotRequired[Nullable[int]] max_uses: NotRequired[Nullable[int]]
user_location: NotRequired[Nullable[AnthropicWebSearchToolUserLocationTypedDict]] user_location: NotRequired[Nullable[AnthropicWebSearchToolUserLocationTypedDict]]
@@ -853,7 +853,7 @@ class ToolWebSearch20250305(BaseModel):
blocked_domains: OptionalNullable[List[str]] = UNSET blocked_domains: OptionalNullable[List[str]] = UNSET
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
max_uses: OptionalNullable[int] = UNSET max_uses: OptionalNullable[int] = UNSET
@@ -905,7 +905,7 @@ class ToolTextEditor20250124TypedDict(TypedDict):
name: NameStrReplaceEditor name: NameStrReplaceEditor
type: TypeTextEditor20250124 type: TypeTextEditor20250124
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
class ToolTextEditor20250124(BaseModel): class ToolTextEditor20250124(BaseModel):
@@ -914,7 +914,7 @@ class ToolTextEditor20250124(BaseModel):
type: TypeTextEditor20250124 type: TypeTextEditor20250124
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
@model_serializer(mode="wrap") @model_serializer(mode="wrap")
def serialize_model(self, handler): def serialize_model(self, handler):
@@ -943,7 +943,7 @@ class ToolBash20250124TypedDict(TypedDict):
name: NameBash name: NameBash
type: TypeBash20250124 type: TypeBash20250124
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
class ToolBash20250124(BaseModel): class ToolBash20250124(BaseModel):
@@ -952,7 +952,7 @@ class ToolBash20250124(BaseModel):
type: TypeBash20250124 type: TypeBash20250124
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
@model_serializer(mode="wrap") @model_serializer(mode="wrap")
def serialize_model(self, handler): def serialize_model(self, handler):
@@ -1033,7 +1033,7 @@ class ToolCustomTypedDict(TypedDict):
input_schema: InputSchemaTypedDict input_schema: InputSchemaTypedDict
name: str name: str
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
description: NotRequired[str] description: NotRequired[str]
type: NotRequired[ToolTypeCustom] type: NotRequired[ToolTypeCustom]
@@ -1044,7 +1044,7 @@ class ToolCustom(BaseModel):
name: str name: str
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
description: Optional[str] = None description: Optional[str] = None
@@ -1113,7 +1113,7 @@ class MessagesRequestTypedDict(TypedDict):
messages: Nullable[List[MessagesMessageParamTypedDict]] messages: Nullable[List[MessagesMessageParamTypedDict]]
model: str model: str
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
context_management: NotRequired[Nullable[ContextManagementTypedDict]] context_management: NotRequired[Nullable[ContextManagementTypedDict]]
fallbacks: NotRequired[Nullable[List[MessagesFallbackParamTypedDict]]] fallbacks: NotRequired[Nullable[List[MessagesFallbackParamTypedDict]]]
r"""Fallback models to try if the primary model fails or refuses, in order. Handled by OpenRouter multi-model routing rather than Anthropic server-side fallbacks; cannot be combined with `models`. Each entry accepts only `model`. Maximum of 3 entries.""" r"""Fallback models to try if the primary model fails or refuses, in order. Handled by OpenRouter multi-model routing rather than Anthropic server-side fallbacks; cannot be combined with `models`. Each entry accepts only `model`. Maximum of 3 entries."""
@@ -1155,7 +1155,7 @@ class MessagesRequest(BaseModel):
model: str model: str
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
context_management: OptionalNullable[ContextManagement] = UNSET context_management: OptionalNullable[ContextManagement] = UNSET
src/openrouter/components/promptcachebreakpoint.py
+2 -2
View File
@@ -10,12 +10,12 @@ PromptCacheBreakpointMode = Literal["explicit",]
class PromptCacheBreakpointTypedDict(TypedDict): class PromptCacheBreakpointTypedDict(TypedDict):
r"""Marks an explicit prompt-cache boundary on this content block. Everything through the block carrying this marker is part of the candidate cached prefix. Only supported by OpenAI GPT-5.6 and newer.""" r"""Marks an explicit prompt-cache boundary on this content block (OpenAI-style). Everything through the block carrying this marker is part of the candidate cached prefix. Supported natively by OpenAI GPT-5.6 and newer; on providers that use Anthropic-style `cache_control`, OpenRouter converts the marker to that format automatically."""
mode: PromptCacheBreakpointMode mode: PromptCacheBreakpointMode
class PromptCacheBreakpoint(BaseModel): class PromptCacheBreakpoint(BaseModel):
r"""Marks an explicit prompt-cache boundary on this content block. Everything through the block carrying this marker is part of the candidate cached prefix. Only supported by OpenAI GPT-5.6 and newer.""" r"""Marks an explicit prompt-cache boundary on this content block (OpenAI-style). Everything through the block carrying this marker is part of the candidate cached prefix. Supported natively by OpenAI GPT-5.6 and newer; on providers that use Anthropic-style `cache_control`, OpenRouter converts the marker to that format automatically."""
mode: PromptCacheBreakpointMode mode: PromptCacheBreakpointMode
src/openrouter/components/responsesrequest.py
+2 -2
View File
@@ -291,7 +291,7 @@ class ResponsesRequestTypedDict(TypedDict):
background: NotRequired[Nullable[bool]] background: NotRequired[Nullable[bool]]
cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict] cache_control: NotRequired[AnthropicCacheControlDirectiveTypedDict]
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
debug: NotRequired[ChatDebugOptionsTypedDict] debug: NotRequired[ChatDebugOptionsTypedDict]
r"""Debug options for inspecting request transformations (streaming only)""" r"""Debug options for inspecting request transformations (streaming only)"""
frequency_penalty: NotRequired[Nullable[float]] frequency_penalty: NotRequired[Nullable[float]]
@@ -351,7 +351,7 @@ class ResponsesRequest(BaseModel):
background: OptionalNullable[bool] = UNSET background: OptionalNullable[bool] = UNSET
cache_control: Optional[AnthropicCacheControlDirective] = None cache_control: Optional[AnthropicCacheControlDirective] = None
r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models.""" r"""Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format."""
debug: Optional[ChatDebugOptions] = None debug: Optional[ChatDebugOptions] = None
r"""Debug options for inspecting request transformations (streaming only)""" r"""Debug options for inspecting request transformations (streaming only)"""
src/openrouter/presets.py
+28 -6
View File
@@ -664,6 +664,10 @@ class Presets(BaseSDK):
] ]
] = None, ] = None,
presence_penalty: OptionalNullable[float] = UNSET, presence_penalty: OptionalNullable[float] = UNSET,
prompt_cache_key: OptionalNullable[str] = UNSET,
prompt_cache_options: OptionalNullable[
Union[components.PromptCacheOptions, components.PromptCacheOptionsTypedDict]
] = UNSET,
provider: OptionalNullable[ provider: OptionalNullable[
Union[ Union[
components.ProviderPreferences, components.ProviderPreferencesTypedDict components.ProviderPreferences, components.ProviderPreferencesTypedDict
@@ -736,7 +740,7 @@ class Presets(BaseSDK):
:param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings. :param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings.
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: Frequency penalty (-2.0 to 2.0) :param frequency_penalty: Frequency penalty (-2.0 to 2.0)
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
@@ -752,6 +756,8 @@ class Presets(BaseSDK):
:param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response. :param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response.
:param plugins: Plugins you want to enable for this request, including their settings. :param plugins: Plugins you want to enable for this request, including their settings.
:param presence_penalty: Presence penalty (-2.0 to 2.0) :param presence_penalty: Presence penalty (-2.0 to 2.0)
:param prompt_cache_key:
:param prompt_cache_options: Request-level prompt-cache controls. `mode: \"explicit\"` disables OpenAI-managed breakpoints so only blocks marked with `prompt_cache_breakpoint` are cached. Only supported by OpenAI GPT-5.6 and newer.
:param provider: When multiple model providers are available, optionally indicate your routing preference. :param provider: When multiple model providers are available, optionally indicate your routing preference.
:param reasoning: Configuration options for reasoning models :param reasoning: Configuration options for reasoning models
:param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ. :param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ.
@@ -825,6 +831,11 @@ class Presets(BaseSDK):
plugins, Optional[List[components.ChatRequestPlugin]] plugins, Optional[List[components.ChatRequestPlugin]]
), ),
presence_penalty=presence_penalty, presence_penalty=presence_penalty,
prompt_cache_key=prompt_cache_key,
prompt_cache_options=utils.get_pydantic_model(
prompt_cache_options,
OptionalNullable[components.PromptCacheOptions],
),
provider=utils.get_pydantic_model( provider=utils.get_pydantic_model(
provider, OptionalNullable[components.ProviderPreferences] provider, OptionalNullable[components.ProviderPreferences]
), ),
@@ -1011,6 +1022,10 @@ class Presets(BaseSDK):
] ]
] = None, ] = None,
presence_penalty: OptionalNullable[float] = UNSET, presence_penalty: OptionalNullable[float] = UNSET,
prompt_cache_key: OptionalNullable[str] = UNSET,
prompt_cache_options: OptionalNullable[
Union[components.PromptCacheOptions, components.PromptCacheOptionsTypedDict]
] = UNSET,
provider: OptionalNullable[ provider: OptionalNullable[
Union[ Union[
components.ProviderPreferences, components.ProviderPreferencesTypedDict components.ProviderPreferences, components.ProviderPreferencesTypedDict
@@ -1083,7 +1098,7 @@ class Presets(BaseSDK):
:param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings. :param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings.
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: Frequency penalty (-2.0 to 2.0) :param frequency_penalty: Frequency penalty (-2.0 to 2.0)
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
@@ -1099,6 +1114,8 @@ class Presets(BaseSDK):
:param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response. :param parallel_tool_calls: Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response.
:param plugins: Plugins you want to enable for this request, including their settings. :param plugins: Plugins you want to enable for this request, including their settings.
:param presence_penalty: Presence penalty (-2.0 to 2.0) :param presence_penalty: Presence penalty (-2.0 to 2.0)
:param prompt_cache_key:
:param prompt_cache_options: Request-level prompt-cache controls. `mode: \"explicit\"` disables OpenAI-managed breakpoints so only blocks marked with `prompt_cache_breakpoint` are cached. Only supported by OpenAI GPT-5.6 and newer.
:param provider: When multiple model providers are available, optionally indicate your routing preference. :param provider: When multiple model providers are available, optionally indicate your routing preference.
:param reasoning: Configuration options for reasoning models :param reasoning: Configuration options for reasoning models
:param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ. :param reasoning_effort: Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ.
@@ -1172,6 +1189,11 @@ class Presets(BaseSDK):
plugins, Optional[List[components.ChatRequestPlugin]] plugins, Optional[List[components.ChatRequestPlugin]]
), ),
presence_penalty=presence_penalty, presence_penalty=presence_penalty,
prompt_cache_key=prompt_cache_key,
prompt_cache_options=utils.get_pydantic_model(
prompt_cache_options,
OptionalNullable[components.PromptCacheOptions],
),
provider=utils.get_pydantic_model( provider=utils.get_pydantic_model(
provider, OptionalNullable[components.ProviderPreferences] provider, OptionalNullable[components.ProviderPreferences]
), ),
@@ -1420,7 +1442,7 @@ class Presets(BaseSDK):
:param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings. :param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings.
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param context_management: :param context_management:
:param fallbacks: Fallback models to try if the primary model fails or refuses, in order. Handled by OpenRouter multi-model routing rather than Anthropic server-side fallbacks; cannot be combined with `models`. Each entry accepts only `model`. Maximum of 3 entries. :param fallbacks: Fallback models to try if the primary model fails or refuses, in order. Handled by OpenRouter multi-model routing rather than Anthropic server-side fallbacks; cannot be combined with `models`. Each entry accepts only `model`. Maximum of 3 entries.
:param max_tokens: :param max_tokens:
@@ -1732,7 +1754,7 @@ class Presets(BaseSDK):
:param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings. :param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings.
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param context_management: :param context_management:
:param fallbacks: Fallback models to try if the primary model fails or refuses, in order. Handled by OpenRouter multi-model routing rather than Anthropic server-side fallbacks; cannot be combined with `models`. Each entry accepts only `model`. Maximum of 3 entries. :param fallbacks: Fallback models to try if the primary model fails or refuses, in order. Handled by OpenRouter multi-model routing rather than Anthropic server-side fallbacks; cannot be combined with `models`. Each entry accepts only `model`. Maximum of 3 entries.
:param max_tokens: :param max_tokens:
@@ -2054,7 +2076,7 @@ class Presets(BaseSDK):
:param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings. :param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings.
:param background: :param background:
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: :param frequency_penalty:
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
@@ -2406,7 +2428,7 @@ class Presets(BaseSDK):
:param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings. :param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings.
:param background: :param background:
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: :param frequency_penalty:
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
src/openrouter/responses.py
+8 -8
View File
@@ -137,7 +137,7 @@ class Responses(BaseSDK):
:param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. :param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility.
:param background: :param background:
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: :param frequency_penalty:
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
@@ -295,7 +295,7 @@ class Responses(BaseSDK):
:param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. :param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility.
:param background: :param background:
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: :param frequency_penalty:
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
@@ -456,7 +456,7 @@ class Responses(BaseSDK):
:param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. :param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility.
:param background: :param background:
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: :param frequency_penalty:
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
@@ -616,7 +616,7 @@ class Responses(BaseSDK):
:param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. :param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility.
:param background: :param background:
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: :param frequency_penalty:
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
@@ -1056,7 +1056,7 @@ class Responses(BaseSDK):
:param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. :param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility.
:param background: :param background:
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: :param frequency_penalty:
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
@@ -1214,7 +1214,7 @@ class Responses(BaseSDK):
:param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. :param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility.
:param background: :param background:
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: :param frequency_penalty:
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
@@ -1375,7 +1375,7 @@ class Responses(BaseSDK):
:param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. :param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility.
:param background: :param background:
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: :param frequency_penalty:
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
@@ -1535,7 +1535,7 @@ class Responses(BaseSDK):
:param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility. :param x_open_router_metadata: Opt-in to surface routing metadata on the response under `openrouter_metadata`. Defaults to `disabled`. The legacy header `X-OpenRouter-Experimental-Metadata` is also accepted for backward compatibility.
:param background: :param background:
:param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. Currently supported for Anthropic Claude models. :param cache_control: Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format.
:param debug: Debug options for inspecting request transformations (streaming only) :param debug: Debug options for inspecting request transformations (streaming only)
:param frequency_penalty: :param frequency_penalty:
:param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. :param image_config: Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details.
uv.lock
Generated
+1 -1
View File
@@ -213,7 +213,7 @@ wheels = [
[[package]] [[package]]
name = "openrouter" name = "openrouter"
version = "0.11.31" version = "0.11.32"
source = { editable = "." } source = { editable = "." }
dependencies = [ dependencies = [
{ name = "httpcore" }, { name = "httpcore" },