Compare commits

..
73 Commits
Author SHA1 Message Date
github-actions[bot]GitHubspeakeasybotspeakeasy-github[bot] <128539517+speakeasy-github[bot]@users.noreply.github.com>
10f2890d75 chore: 🐝 Update SDK - Generate (spec change merged) 0.10.7 (#378)
Co-authored-by: speakeasybot <bot@speakeasyapi.dev>
Co-authored-by: speakeasy-github[bot] <128539517+speakeasy-github[bot]@users.noreply.github.com>
2026-06-26 23:48:17 +00:00
ad1cfda1f4 chore: update OpenAPI spec from monorepo (#377)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-26 23:42:21 +00:00
github-actions[bot]GitHubspeakeasybotspeakeasy-github[bot] <128539517+speakeasy-github[bot]@users.noreply.github.com>
21020fb334 chore: 🐝 Update SDK - Generate (spec change merged) 0.10.6 (#376)
Co-authored-by: speakeasybot <bot@speakeasyapi.dev>
Co-authored-by: speakeasy-github[bot] <128539517+speakeasy-github[bot]@users.noreply.github.com>
2026-06-26 18:23:57 +00:00
a243d78d32 chore: update OpenAPI spec from monorepo (#375)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-26 18:17:47 +00:00
github-actions[bot]GitHubspeakeasybotspeakeasy-github[bot] <128539517+speakeasy-github[bot]@users.noreply.github.com>
66512fa808 chore: 🐝 Update SDK - Generate (spec change merged) 0.10.5 (#374)
Co-authored-by: speakeasybot <bot@speakeasyapi.dev>
Co-authored-by: speakeasy-github[bot] <128539517+speakeasy-github[bot]@users.noreply.github.com>
2026-06-26 14:41:46 +00:00
7efd03fb3b chore: update OpenAPI spec from monorepo (#373)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-26 14:35:36 +00:00
github-actions[bot]GitHubspeakeasybotspeakeasy-github[bot] <128539517+speakeasy-github[bot]@users.noreply.github.com>
e1b89f5aaf chore: 🐝 Update SDK - Generate (spec change merged) 0.10.4 (#372)
Co-authored-by: speakeasybot <bot@speakeasyapi.dev>
Co-authored-by: speakeasy-github[bot] <128539517+speakeasy-github[bot]@users.noreply.github.com>
2026-06-26 12:10:48 +00:00
32dd522d5a chore: update OpenAPI spec from monorepo (#371)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-26 12:04:34 +00:00
github-actions[bot]GitHubspeakeasybotspeakeasy-github[bot] <128539517+speakeasy-github[bot]@users.noreply.github.com>
eeb0373033 chore: 🐝 Update SDK - Generate (spec change merged) 0.10.3 (#370)
Co-authored-by: speakeasybot <bot@speakeasyapi.dev>
Co-authored-by: speakeasy-github[bot] <128539517+speakeasy-github[bot]@users.noreply.github.com>
2026-06-26 00:20:39 +00:00
facade7d00 chore: update OpenAPI spec from monorepo (#369)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-26 00:14:27 +00:00
github-actions[bot]GitHubspeakeasybotspeakeasy-github[bot] <128539517+speakeasy-github[bot]@users.noreply.github.com>
08e796401a chore: 🐝 Update SDK - Generate (spec change merged) 0.10.2 (#368)
Co-authored-by: speakeasybot <bot@speakeasyapi.dev>
Co-authored-by: speakeasy-github[bot] <128539517+speakeasy-github[bot]@users.noreply.github.com>
2026-06-25 22:33:28 +00:00
f8e8ab669d chore: update OpenAPI spec from monorepo (#367)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-25 22:27:08 +00:00
github-actions[bot]GitHubspeakeasybotspeakeasy-github[bot] <128539517+speakeasy-github[bot]@users.noreply.github.com>
5e0b1a2b69 chore: 🐝 Update SDK - Generate 0.10.1 (#366)
Co-authored-by: speakeasybot <bot@speakeasyapi.dev>
Co-authored-by: speakeasy-github[bot] <128539517+speakeasy-github[bot]@users.noreply.github.com>
2026-06-25 22:01:36 +00:00
Christine ChenandGitHub 667941aaa8 fix: chain Speakeasy auto-merge into Generate run (#364) 2026-06-25 17:53:58 -04:00
Christine ChenandGitHub 011fb81cc3 feat: add PR validation workflow (#365) 2026-06-25 17:53:42 -04:00
Christine Chen 5c06b57cea feat: add PR validation workflow
The Python SDK had no pull_request-triggered checks, so regen PRs
(speakeasy-sdk-regen-*) carried an empty status-check rollup and the
auto-merge flow would merge them with zero validation. The TypeScript
SDK validates every regen PR via pr-validation.yaml; Python had no
equivalent.

Add a single-file PR Validation workflow that builds the package and
runs the SDK's existing quality gates (mypy, pyright, pylint) via uv.
No tests exist in this repo yet, so no API-key secret is needed, which
also lets the workflow run on fork PRs. paths-ignore on
.speakeasy/in.openapi.yaml mirrors TS so pure spec-bump PRs don't
double-trigger.

This registers a `validate` check that the auto-merge wait_for_checks
step will gate on once both land.
2026-06-25 17:31:33 -04:00
Christine Chen 0680f01bf7 fix: proceed on checkless regen PRs in wait_for_checks
Speakeasy regen PRs are opened with the workflow GITHUB_TOKEN, which by
design does not trigger downstream on: pull_request runs, so they carry
zero status checks. main has no required checks either, so gh pr merge
--auto no-ops and the direct-merge fallback calls wait_for_checks.

The empty-rollup arm only `continue`d, so TOTAL stayed 0 until the 600s
timeout and the function exit 1'd — failing auto-merge on every regen PR.
Add a 60s grace window for checks to register, then treat a persistently
empty rollup as "no checks to wait for" and return 0 so the direct squash
merge proceeds. Matches the fix already shipped in go-sdk #318.
2026-06-25 16:20:53 -04:00
Christine Chen 186c1075ce fix: chain Speakeasy auto-merge into Generate run
The auto-merge workflow triggered on `pull_request: [labeled, opened]`.
When github-actions[bot] opened the regen PR, GitHub's recursion guard
refused to run a workflow triggered by another workflow's GITHUB_TOKEN,
so the run landed as action_required with zero jobs executed. Regen PRs
piled up unmerged and sdk_publish never fired (PyPI stuck at 0.10.0).

Port the TypeScript SDK's architecture: make auto-merge a workflow_call
reusable workflow chained inside the Generate run (generate ->
resolve-branch -> auto-merge). No separate pull_request-triggered run
exists, so there is nothing for GitHub to gate.

- auto-merge-speakeasy-pr.yaml: workflow_call inputs (run_started_at,
  branch_name) + App-token secrets; resolve PR via branch/timestamp;
  .author.is_bot check; wait_for_checks before direct merge; step
  summaries. Keeps the SHA-pinned create-github-app-token.
- sdk_generation.yaml / sdk_generation_for_spec_change.yaml: add
  concurrency group, resolve-branch + auto-merge jobs.
2026-06-25 16:08:16 -04:00
b161a5c288 chore: update OpenAPI spec from monorepo (#362)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-24 22:33:22 +00:00
4f018c91d7 chore: update OpenAPI spec from monorepo (#361)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-24 20:57:38 +00:00
63949d45e2 chore: update OpenAPI spec from monorepo (#360)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-24 17:13:41 +00:00
399c1d0e02 chore: update OpenAPI spec from monorepo (#359)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-24 02:53:25 +00:00
89717dca56 chore: update OpenAPI spec from monorepo (#356)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-24 01:01:29 +00:00
Christine ChenandGitHub c8bf74a516 fix: use GitHub App token for auto-merge workflow (#355) 2026-06-23 19:55:41 -04:00
Christine Chen 4e551e432f fix: pin create-github-app-token to immutable SHA
Pin actions/create-github-app-token to v3.2.0 commit SHA (bcd2ba49...) instead of mutable @v3 tag to prevent supply chain attacks where a compromised account could force-push malicious code to the tag. This eliminates the risk of secret leakage and malicious code execution in CI workflows.
2026-06-23 18:14:27 -04:00
Christine Chen 60add6da49 fix: use GitHub App token for auto-merge workflow
Generate GitHub App token using GH_DOCS_SYNC_APP_ID and GH_DOCS_SYNC_APP_PRIVATE_KEY secrets. Use the app token instead of GITHUB_TOKEN for both close-superseded and auto-merge steps to ensure downstream workflow triggers (sdk_publish.yaml) after merge. This aligns token usage with typescript-sdk and go-sdk implementations.
2026-06-23 17:59:16 -04:00
ddcadce542 chore: update OpenAPI spec from monorepo (#354)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-23 21:46:18 +00:00
5cefd3a108 chore: update OpenAPI spec from monorepo (#353)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-23 19:02:28 +00:00
713dd5c71a chore: update OpenAPI spec from monorepo (#351)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-23 14:08:18 +00:00
678d5aacc5 chore: update OpenAPI spec from monorepo (#350)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-22 22:11:11 +00:00
b3f4ab7452 chore: update OpenAPI spec from monorepo (#348)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-22 10:28:13 -07:00
Christine ChenandGitHub 7b4b9feaa5 fix: update auto-merge workflow triggers and bot username (#345) 2026-06-22 12:32:12 -04:00
Christine Chen add423f762 fix: add updated workflow with corrected bot username and event handling
- Add 'opened' trigger to catch PRs when bot immediately applies label
- Fix bot username to github-actions[bot] (verified via API and commit history)
- Restructure condition to handle both 'labeled' and 'opened' events
  * 'labeled': uses github.event.label.name
  * 'opened': uses github.event.pull_request.labels.*.name
  * Prevents 'opened' event from failing due to missing github.event.label
This enables the auto-merge workflow to work with Speakeasy SDK generation PRs.
2026-06-22 11:59:27 -04:00
Christine Chen c34ce15fa2 fix: update auto-merge workflow triggers and bot username
- Add 'opened' trigger to catch PRs when bot immediately applies label
- Fix bot username from github-actions[bot] to app/github-actions
This enables the auto-merge workflow to work with Speakeasy SDK generation PRs.
2026-06-22 11:20:24 -04:00
8e637041c4 chore: update OpenAPI spec from monorepo (#343)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-22 08:00:49 -07:00
146cec3c08 chore: update OpenAPI spec from monorepo (#341)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-19 09:03:24 -07:00
14a62e3451 chore: update OpenAPI spec from monorepo (#339)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-18 16:24:52 -07:00
7e29e408af chore: update OpenAPI spec from monorepo (#333)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-18 10:20:41 -07:00
Christine ChenandGitHub 30bc131b97 feat: auto-merge Speakeasy SDK generation PRs (#332) 2026-06-18 11:05:15 -04:00
Christine Chen 1b0118c365 fix: address review feedback on auto-merge workflow
Pipe gh pr list output through jq instead of passing --arg to gh --jq,
and add a repo-level concurrency group to prevent cross-PR merge races.
2026-06-17 16:54:40 -04:00
Christine Chen cda73e5ebf feat: auto-merge Speakeasy SDK generation PRs
Add a workflow that merges Speakeasy codegen PRs when the bot applies a
semver label, closing superseded regen PRs first. This unblocks
sdk_publish.yaml after mode:pr generation without manual intervention.
2026-06-17 16:28:01 -04:00
b0eda78d3f chore: update OpenAPI spec from monorepo (#330)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-17 12:53:02 -07:00
Christine ChenandGitHub bf682bdb62 fix: use PR mode for spec-change SDK generation (#329) 2026-06-17 15:07:36 -04:00
Christine Chen b322622117 fix: use PR mode for spec-change SDK generation
Switch sdk_generation_for_spec_change.yaml from direct to pr mode so
Speakeasy opens a PR instead of pushing to main. This avoids branch
protection failures and routes publishing through sdk_publish.yaml.
2026-06-17 10:37:08 -04:00
Matt AppersonandGitHub 415450e785 chore: 🐝 Update SDK - Generate 0.10.0 (#317) 2026-06-17 09:48:04 -04:00
speakeasy-github[bot]andspeakeasy-bot 1d488f0117 empty commit to trigger [run-tests] workflow 2026-06-17 01:01:27 +00:00
speakeasybot 4e789208b4 ## Python SDK Changes:
* `open_router.beta.responses.send()`: 
  *  `request` **Changed** **Breaking** ⚠️
  *  `response` **Changed** **Breaking** ⚠️
* `open_router.presets.create_presets_responses()`:  `request` **Changed** **Breaking** ⚠️
* `open_router.presets.create_presets_chat_completions()`:  `request` **Changed** **Breaking** ⚠️
* `open_router.chat.send()`:  `request` **Changed** **Breaking** ⚠️
* `open_router.workspaces.set_budget()`: **Added**
* `open_router.o_auth.create_auth_code()`: 
  *  `request.workspace_id` **Added**
  *  `error.status[403]` **Added**
* `open_router.files.download()`: **Added**
* `open_router.models.get()`: **Added**
* `open_router.workspaces.list_budgets()`: **Added**
* `open_router.workspaces.delete_budget()`: **Added**
* `open_router.datasets.get_benchmarks_artificial_analysis()`: **Added**
* `open_router.beta.analytics.query_analytics()`:  `response.data.warnings` **Added**
* `open_router.files.delete()`: **Added**
* `open_router.files.retrieve()`: **Added**
* `open_router.files.upload()`: **Added**
* `open_router.embeddings.list_models()`:  `response.data.[].benchmarks` **Added**
* `open_router.models.list()`: 
  *  `request` **Changed**
  *  `response.data.[].benchmarks` **Added**
* `open_router.models.list_for_user()`:  `response.data.[].benchmarks` **Added**
* `open_router.files.list()`: **Added**
* `open_router.presets.create_presets_messages()`:  `request` **Changed**
* `open_router.datasets.get_benchmarks_design_arena()`: **Added**
2026-06-17 01:01:09 +00:00
cdd4dc7302 chore: update OpenAPI spec from monorepo (#328)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-16 16:01:05 -07:00
Christine ChenandGitHub 60708e1e8b docs: remove beta labels from Python SDK README (#327) 2026-06-16 18:58:58 -04:00
Christine Chen 5a324f99ae docs: remove beta labels from Python SDK README
The Python SDK is no longer in beta. Drop the (Beta) title suffix and
remove the Maturity section disclaimer from README and README-PYPI.

SDK-464
2026-06-16 18:09:12 -04:00
d5778dfafb chore: update OpenAPI spec from monorepo (#326)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-16 14:36:02 -07:00
68c251830c chore: update OpenAPI spec from monorepo (#325)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-16 12:52:37 -07:00
2aa2d73d2d chore: update OpenAPI spec from monorepo (#324)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-16 10:37:39 -07:00
58bd390fbd chore: update OpenAPI spec from monorepo (#323)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-16 00:39:20 -07:00
2707901239 chore: update OpenAPI spec from monorepo (#322)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-15 14:21:32 -07:00
a5ed2de8a8 chore: update OpenAPI spec from monorepo (#321)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-15 10:29:12 -07:00
robert-j-yandGitHub 21893dda58 ci: regenerate SDK directly when merged spec changes land (#315) 2026-06-15 09:52:12 -07:00
758a2f3d49 chore: update OpenAPI spec from monorepo (#320)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-15 09:28:46 -07:00
0331f14f19 chore: update OpenAPI spec from monorepo (#319)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-15 06:48:08 -07:00
138b131a35 chore: update OpenAPI spec from monorepo (#318)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-12 22:08:18 -07:00
Christine ChenandGitHub 55c0d5c7fd docs: update outdated model counts in SDK documentation (#313) 2026-06-12 17:37:19 -04:00
e98a7a8260 chore: update OpenAPI spec from monorepo (#316)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-12 13:38:51 -07:00
Christine Chen 7e7787cb23 docs: update outdated model counts in SDK documentation
Refresh README and overview copy to reflect 400+ available models instead of 300+.
2026-06-12 15:47:18 -04:00
c4b508f045 chore: update OpenAPI spec from monorepo (#314)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-12 12:44:30 -07:00
52e7853d21 chore: update OpenAPI spec from monorepo (#312)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-12 10:46:58 -07:00
d43061f722 chore: update OpenAPI spec from monorepo (#311)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-12 10:09:55 -07:00
7b10fd47d8 chore: update OpenAPI spec from monorepo (#310)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-12 07:28:06 -07:00
95f2c87bfa chore: update OpenAPI spec from monorepo (#309)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-12 06:51:21 -07:00
c756fc51e0 chore: update OpenAPI spec from monorepo (#308)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-11 17:50:16 -07:00
386db75669 chore: update OpenAPI spec from monorepo (#307)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-11 13:27:05 -07:00
e3dc2e6f4c chore: update OpenAPI spec from monorepo (#306)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-11 12:10:20 -07:00
70c8757fc0 chore: update OpenAPI spec from monorepo (#305)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-11 11:10:32 -07:00
b794150ed9 chore: update OpenAPI spec from monorepo (#304)
Co-authored-by: OpenRouter SDK Bot <sdk-bot@openrouter.ai>
2026-06-11 09:18:46 -07:00
171 changed files with 21389 additions and 1094 deletions
@@ -0,0 +1,225 @@
name: Auto-merge Speakeasy PR
# Called as a follow-up job after Speakeasy Generate (mode: pr). Merges the
# regen PR opened/updated in the same workflow run so sdk_publish.yaml can run.
# Prefer branch_name from Generate logs; fall back to run_started_at heuristics.
# Uses GH_DOCS_SYNC GitHub App token (not a PAT) so merges trigger downstream push.
on:
workflow_call:
inputs:
run_started_at:
description: ISO timestamp when the parent Generate workflow run started
required: true
type: string
branch_name:
description: Speakeasy regen branch from the parent Generate run (optional)
required: false
type: string
default: ""
secrets:
GH_DOCS_SYNC_APP_ID:
required: true
GH_DOCS_SYNC_APP_PRIVATE_KEY:
required: true
permissions:
contents: write
pull-requests: write
concurrency:
group: auto-merge-speakeasy
cancel-in-progress: false
jobs:
auto-merge:
runs-on: ubuntu-latest
steps:
- name: Generate GitHub App token
id: app-token
# actions/create-github-app-token@v3.2.0 (immutable SHA)
uses: actions/create-github-app-token@bcd2ba49218906704ab6c1aa796996da409d3eb1
with:
app-id: ${{ secrets.GH_DOCS_SYNC_APP_ID }}
private-key: ${{ secrets.GH_DOCS_SYNC_APP_PRIVATE_KEY }}
- name: Resolve Speakeasy regen PR
id: pr
env:
GH_TOKEN: ${{ steps.app-token.outputs.token }}
run: |
set -euo pipefail
REPO="${{ github.repository }}"
RUN_STARTED="${{ inputs.run_started_at }}"
BRANCH="${{ inputs.branch_name }}"
PR_JSON=$(gh pr list \
--repo "$REPO" \
--state open \
--search "head:speakeasy-sdk-regen-" \
--limit 500 \
--json number,updatedAt,title,labels,author,headRefName)
PR_NUM=""
if [ -n "$BRANCH" ]; then
CANDIDATE=$(echo "$PR_JSON" | jq -r --arg branch "$BRANCH" \
'[.[] | select(.headRefName == $branch)] | .[0] // empty')
if [ -n "$CANDIDATE" ] && echo "$CANDIDATE" | jq -e '
(.headRefName | startswith("speakeasy-sdk-regen-"))
and (.title | contains("🐝 Update SDK"))
and ([.labels[].name] | any(. == "patch" or . == "minor" or . == "major"))
and (.author.is_bot == true)
' > /dev/null; then
PR_NUM=$(echo "$CANDIDATE" | jq -r '.number')
echo "Resolved Speakeasy regen PR #$PR_NUM from branch $BRANCH"
else
echo "::warning::branch_name=$BRANCH did not match a valid open regen PR — falling back to run_started_at"
fi
fi
if [ -z "$PR_NUM" ]; then
PR_NUM=$(echo "$PR_JSON" | jq -r --arg started "$RUN_STARTED" '
[.[]
| select(.headRefName | startswith("speakeasy-sdk-regen-"))
| select(.title | contains("🐝 Update SDK"))
| select([.labels[].name] | any(. == "patch" or . == "minor" or . == "major"))
| select(.author.is_bot == true)
| select(.updatedAt >= $started)
] | sort_by(.updatedAt) | last | .number // empty')
fi
if [ -z "$PR_NUM" ]; then
echo "No Speakeasy regen PR for this Generate run — skipping auto-merge"
echo "::notice title=auto-merge-skipped::No qualifying Speakeasy regen PR to merge"
{
echo "## Auto-merge skipped"
echo "No qualifying \`speakeasy-sdk-regen-*\` PR was found."
echo "- branch_name input: \`${BRANCH:-<empty>}\`"
echo "- run_started_at: \`$RUN_STARTED\`"
} >> "$GITHUB_STEP_SUMMARY"
echo "pr_num=" >> "$GITHUB_OUTPUT"
exit 0
fi
echo "pr_num=$PR_NUM" >> "$GITHUB_OUTPUT"
echo "$PR_JSON" > /tmp/speakeasy_pr_json.json
- name: Close superseded Speakeasy PRs
if: steps.pr.outputs.pr_num != ''
env:
GH_TOKEN: ${{ steps.app-token.outputs.token }}
run: |
set -euo pipefail
CURRENT_PR="${{ steps.pr.outputs.pr_num }}"
PR_JSON=$(cat /tmp/speakeasy_pr_json.json)
PRIOR_JSON=$(echo "$PR_JSON" | jq -r --arg current_pr "$CURRENT_PR" \
'[.[].number | select(tostring != $current_pr)] | .[]')
if [ -z "$PRIOR_JSON" ]; then
echo "No superseded Speakeasy PRs to close"
exit 0
fi
mapfile -t PRIOR <<< "$PRIOR_JSON"
echo "Closing ${#PRIOR[@]} superseded Speakeasy PR(s)"
for N in "${PRIOR[@]}"; do
gh pr close "$N" --repo "${{ github.repository }}" --delete-branch \
--comment "Superseded by newer Speakeasy SDK generation PR #${CURRENT_PR}" \
|| echo "::warning::Failed to close PR #$N (continuing)"
sleep 1
done
- name: Auto-merge Speakeasy PR
if: steps.pr.outputs.pr_num != ''
env:
GH_TOKEN: ${{ steps.app-token.outputs.token }}
run: |
set -euo pipefail
PR_NUM="${{ steps.pr.outputs.pr_num }}"
REPO="${{ github.repository }}"
wait_for_checks() {
local pr=$1
local timeout=600
local interval=15
local elapsed=0
echo "Waiting for PR #$pr checks (timeout ${timeout}s)..."
while [ "$elapsed" -lt "$timeout" ]; do
TOTAL=$(gh pr view "$pr" --repo "$REPO" --json statusCheckRollup --jq \
'.statusCheckRollup | length')
PENDING=$(gh pr view "$pr" --repo "$REPO" --json statusCheckRollup --jq \
'[.statusCheckRollup[]? | select(.status != "COMPLETED")] | length')
FAILED=$(gh pr view "$pr" --repo "$REPO" --json statusCheckRollup --jq \
'[.statusCheckRollup[]? | select(.status == "COMPLETED") | select(.conclusion != "SUCCESS" and .conclusion != "SKIPPED" and .conclusion != "NEUTRAL")] | length')
if [ "$TOTAL" -eq 0 ]; then
# GITHUB_TOKEN-authored regen PRs get no pull_request checks at all,
# so TOTAL stays 0 forever. Wait a short grace window in case checks
# are merely slow to register, then proceed rather than time out.
if [ "$elapsed" -ge 60 ]; then
echo "No checks registered after ${elapsed}s — proceeding (checkless PR)"
return 0
fi
echo "Checks not yet registered — waiting ${interval}s..."
sleep "$interval"
elapsed=$((elapsed + interval))
continue
fi
if [ "$PENDING" -eq 0 ]; then
if [ "$FAILED" -gt 0 ]; then
echo "::error::PR #$pr has failing checks — refusing direct merge"
gh pr checks "$pr" --repo "$REPO" || true
exit 1
fi
echo "All checks completed"
return 0
fi
echo "Checks pending ($PENDING) — waiting ${interval}s..."
sleep "$interval"
elapsed=$((elapsed + interval))
done
echo "::error::Timed out waiting for PR #$pr checks"
exit 1
}
STATE=$(gh pr view "$PR_NUM" --repo "$REPO" --json state --jq .state)
if [ "$STATE" != "OPEN" ]; then
echo "PR #$PR_NUM is $STATE — skipping merge"
exit 0
fi
MERGE_METHOD=""
# Prefer GitHub auto-merge when required checks exist; otherwise
# squash-merge directly after checks pass (sdk-release-prs.yaml pattern).
AUTO_MERGE_NOOP_PATTERN='is in clean status|protected branch rules|Branch does not have required protected branch rules'
if gh pr merge "$PR_NUM" --repo "$REPO" --squash --auto --delete-branch 2> /tmp/gh-merge.err; then
MERGE_METHOD="auto-merge queued"
echo "Auto-merge enabled for Speakeasy PR #$PR_NUM"
elif grep -qiE "$AUTO_MERGE_NOOP_PATTERN" /tmp/gh-merge.err; then
echo "Auto-merge not applicable — waiting for checks, then merging PR #$PR_NUM directly"
cat /tmp/gh-merge.err >&2
wait_for_checks "$PR_NUM"
gh pr merge "$PR_NUM" --repo "$REPO" --squash --delete-branch
MERGE_METHOD="direct squash (checks passed)"
else
echo "::error::Failed to enable auto-merge on Speakeasy PR #$PR_NUM"
cat /tmp/gh-merge.err >&2
exit 1
fi
echo "::notice title=auto-merge-pr::Merged Speakeasy PR #$PR_NUM ($MERGE_METHOD)"
{
echo "## Auto-merge complete"
echo "- PR: #$PR_NUM"
echo "- Method: $MERGE_METHOD"
echo "- Token: GH_DOCS_SYNC GitHub App"
} >> "$GITHUB_STEP_SUMMARY"
+37
View File
@@ -0,0 +1,37 @@
name: PR Validation
on:
pull_request:
paths-ignore:
- .speakeasy/in.openapi.yaml
concurrency:
group: pr-validation-${{ github.event.pull_request.number }}
cancel-in-progress: true
permissions:
contents: read
jobs:
validate:
runs-on: ubuntu-latest
steps:
- name: Checkout code
uses: actions/checkout@v5
- name: Install uv
uses: astral-sh/setup-uv@v5
with:
enable-cache: true
- name: Build SDK
run: uv build
- name: Type check (mypy)
run: uv run --group dev mypy src
- name: Type check (pyright)
run: uv run --group dev pyright src
- name: Lint (pylint)
run: uv run --group dev pylint src --rcfile pylintrc
+69
View File
@@ -21,6 +21,11 @@ permissions:
types:
- labeled
- unlabeled
concurrency:
group: speakeasy-generate
cancel-in-progress: false
jobs:
generate:
uses: speakeasy-api/sdk-generation-action/.github/workflows/workflow-executor.yaml@v15
@@ -32,3 +37,67 @@ jobs:
github_access_token: ${{ secrets.GITHUB_TOKEN }}
pypi_token: ${{ secrets.PYPI_TOKEN }}
speakeasy_api_key: ${{ secrets.SPEAKEASY_API_KEY }}
resolve-branch:
needs: generate
if: ${{ !cancelled() && needs.generate.result == 'success' }}
runs-on: ubuntu-latest
outputs:
branch_name: ${{ steps.branch.outputs.branch_name }}
run_started_at: ${{ steps.branch.outputs.run_started_at }}
steps:
- name: Extract branch from Generate logs
id: branch
env:
GH_TOKEN: ${{ github.token }}
run: |
set -euo pipefail
REPO="${{ github.repository }}"
RUN_ID="${{ github.run_id }}"
RUN_STARTED="${{ github.run_started_at }}"
if [ -z "$RUN_STARTED" ]; then
RUN_STARTED=$(gh run view "$RUN_ID" --repo "$REPO" --json startedAt --jq '.startedAt')
echo "Resolved run_started_at from API: $RUN_STARTED"
fi
echo "run_started_at=$RUN_STARTED" >> "$GITHUB_OUTPUT"
# Full-run logs are unavailable while the workflow is in progress;
# fetch logs from the completed generate job instead.
JOB_ID=$(gh run view "$RUN_ID" --repo "$REPO" --json jobs --jq \
'[.jobs[] | select(.name == "generate / Generate Target")] | .[0].databaseId // empty')
BRANCH=""
if [ -z "$JOB_ID" ]; then
echo "::warning::Could not find generate job id — auto-merge will use run_started_at fallback"
else
MAX_ATTEMPTS=12
INTERVAL=10
for attempt in $(seq 1 "$MAX_ATTEMPTS"); do
BRANCH=$(gh run view "$RUN_ID" --repo "$REPO" --job "$JOB_ID" --log 2>/dev/null \
| grep -oE 'branch_name=speakeasy-sdk-regen-[0-9]+' | tail -1 | cut -d= -f2 || true)
if [ -n "$BRANCH" ]; then
echo "Extracted branch_name=$BRANCH (attempt $attempt)"
break
fi
if [ "$attempt" -lt "$MAX_ATTEMPTS" ]; then
echo "Job logs not ready — waiting ${INTERVAL}s (attempt $attempt/$MAX_ATTEMPTS)..."
sleep "$INTERVAL"
fi
done
if [ -z "$BRANCH" ]; then
echo "::warning::Could not extract branch_name from Generate logs after ${MAX_ATTEMPTS} attempts — auto-merge will use run_started_at fallback"
fi
fi
echo "branch_name=$BRANCH" >> "$GITHUB_OUTPUT"
auto-merge:
needs: [generate, resolve-branch]
if: ${{ !cancelled() && needs.generate.result == 'success' }}
uses: ./.github/workflows/auto-merge-speakeasy-pr.yaml
with:
run_started_at: ${{ needs.resolve-branch.outputs.run_started_at }}
branch_name: ${{ needs.resolve-branch.outputs.branch_name }}
secrets:
GH_DOCS_SYNC_APP_ID: ${{ secrets.GH_DOCS_SYNC_APP_ID }}
GH_DOCS_SYNC_APP_PRIVATE_KEY: ${{ secrets.GH_DOCS_SYNC_APP_PRIVATE_KEY }}
@@ -0,0 +1,104 @@
name: Generate (spec change merged)
permissions:
checks: write
contents: write
pull-requests: write
statuses: write
id-token: write
"on":
workflow_dispatch:
inputs:
force:
description: Force generation of SDKs
type: boolean
default: false
set_version:
description: optionally set a specific SDK version
type: string
pull_request:
types: [closed]
branches:
- main
paths:
- .speakeasy/in.openapi.yaml
concurrency:
group: speakeasy-generate-spec-change
cancel-in-progress: false
jobs:
generate:
if: github.event.pull_request.merged == true || github.event_name == 'workflow_dispatch'
uses: speakeasy-api/sdk-generation-action/.github/workflows/workflow-executor.yaml@v15
with:
force: ${{ github.event.inputs.force }}
mode: pr
set_version: ${{ github.event.inputs.set_version }}
secrets:
github_access_token: ${{ secrets.GITHUB_TOKEN }}
pypi_token: ${{ secrets.PYPI_TOKEN }}
speakeasy_api_key: ${{ secrets.SPEAKEASY_API_KEY }}
resolve-branch:
needs: generate
if: ${{ !cancelled() && needs.generate.result == 'success' }}
runs-on: ubuntu-latest
outputs:
branch_name: ${{ steps.branch.outputs.branch_name }}
run_started_at: ${{ steps.branch.outputs.run_started_at }}
steps:
- name: Extract branch from Generate logs
id: branch
env:
GH_TOKEN: ${{ github.token }}
run: |
set -euo pipefail
REPO="${{ github.repository }}"
RUN_ID="${{ github.run_id }}"
RUN_STARTED="${{ github.run_started_at }}"
if [ -z "$RUN_STARTED" ]; then
RUN_STARTED=$(gh run view "$RUN_ID" --repo "$REPO" --json startedAt --jq '.startedAt')
echo "Resolved run_started_at from API: $RUN_STARTED"
fi
echo "run_started_at=$RUN_STARTED" >> "$GITHUB_OUTPUT"
# Full-run logs are unavailable while the workflow is in progress;
# fetch logs from the completed generate job instead.
JOB_ID=$(gh run view "$RUN_ID" --repo "$REPO" --json jobs --jq \
'[.jobs[] | select(.name == "generate / Generate Target")] | .[0].databaseId // empty')
BRANCH=""
if [ -z "$JOB_ID" ]; then
echo "::warning::Could not find generate job id — auto-merge will use run_started_at fallback"
else
MAX_ATTEMPTS=12
INTERVAL=10
for attempt in $(seq 1 "$MAX_ATTEMPTS"); do
BRANCH=$(gh run view "$RUN_ID" --repo "$REPO" --job "$JOB_ID" --log 2>/dev/null \
| grep -oE 'branch_name=speakeasy-sdk-regen-[0-9]+' | tail -1 | cut -d= -f2 || true)
if [ -n "$BRANCH" ]; then
echo "Extracted branch_name=$BRANCH (attempt $attempt)"
break
fi
if [ "$attempt" -lt "$MAX_ATTEMPTS" ]; then
echo "Job logs not ready — waiting ${INTERVAL}s (attempt $attempt/$MAX_ATTEMPTS)..."
sleep "$INTERVAL"
fi
done
if [ -z "$BRANCH" ]; then
echo "::warning::Could not extract branch_name from Generate logs after ${MAX_ATTEMPTS} attempts — auto-merge will use run_started_at fallback"
fi
fi
echo "branch_name=$BRANCH" >> "$GITHUB_OUTPUT"
auto-merge:
needs: [generate, resolve-branch]
if: ${{ !cancelled() && needs.generate.result == 'success' }}
uses: ./.github/workflows/auto-merge-speakeasy-pr.yaml
with:
run_started_at: ${{ needs.resolve-branch.outputs.run_started_at }}
branch_name: ${{ needs.resolve-branch.outputs.branch_name }}
secrets:
GH_DOCS_SYNC_APP_ID: ${{ secrets.GH_DOCS_SYNC_APP_ID }}
GH_DOCS_SYNC_APP_PRIVATE_KEY: ${{ secrets.GH_DOCS_SYNC_APP_PRIVATE_KEY }}
+1472 -315
View File
File diff suppressed because one or more lines are too long
+1 -1
View File
@@ -32,7 +32,7 @@ generation:
skipResponseBodyAssertions: false
preApplyUnionDiscriminators: true
python:
version: 0.9.2
version: 0.10.7
additionalDependencies:
dev: {}
main: {}
+3661 -177
View File
File diff suppressed because it is too large Load Diff
+3579 -140
View File
File diff suppressed because it is too large Load Diff
+6 -6
View File
@@ -2,20 +2,20 @@ speakeasyVersion: 1.680.0
sources:
OpenRouter API:
sourceNamespace: open-router-chat-completions-api
sourceRevisionDigest: sha256:08108f43d4fba1406b4c8f624792054965c95a14819f442262663cb25ce6cc25
sourceBlobDigest: sha256:47fc5915e6a9ff29fca019515d6491adce926407ebb47cee480feb77cfb68f03
sourceRevisionDigest: sha256:c668cc885c117987e3c561af619ec3ad21dbb47bffdc491f06d93b8af0c122d3
sourceBlobDigest: sha256:da0dd6040f1f11463b596a1425a6ed5c6f86b5b5e3de2b761cb426f3dcc1660c
tags:
- latest
- speakeasy-sdk-regen-1776991284
- speakeasy-sdk-regen-1782517441
- 1.0.0
targets:
open-router:
source: OpenRouter API
sourceNamespace: open-router-chat-completions-api
sourceRevisionDigest: sha256:08108f43d4fba1406b4c8f624792054965c95a14819f442262663cb25ce6cc25
sourceBlobDigest: sha256:47fc5915e6a9ff29fca019515d6491adce926407ebb47cee480feb77cfb68f03
sourceRevisionDigest: sha256:c668cc885c117987e3c561af619ec3ad21dbb47bffdc491f06d93b8af0c122d3
sourceBlobDigest: sha256:da0dd6040f1f11463b596a1425a6ed5c6f86b5b5e3de2b761cb426f3dcc1660c
codeSamplesNamespace: open-router-python-code-samples
codeSamplesRevisionDigest: sha256:3c7b0aef799130660fd4017ab9113e6fd3e06a80cff6a540043b9b34dc22facf
codeSamplesRevisionDigest: sha256:163ae66544ac5b5162821baae4129dd200441bac24bdacca44316db38a9b178f
workflow:
workflowVersion: 1.0.0
speakeasyVersion: 1.680.0
+1 -1
View File
@@ -1,6 +1,6 @@
# OpenRouter Python SDK
The OpenRouter Python SDK is a type-safe toolkit for building AI applications with access to 300+ language models through a unified API.
The OpenRouter Python SDK is a type-safe toolkit for building AI applications with access to 400+ language models through a unified API.
## Why use the OpenRouter SDK?
+36 -9
View File
@@ -1,6 +1,6 @@
# OpenRouter SDK (Beta)
# OpenRouter SDK
The [OpenRouter SDK](https://openrouter.ai/docs/sdks/python) is a Python toolkit designed to help you build AI-powered features and solutions. Giving you easy access to over 300 models across providers in an easy and type-safe way.
The [OpenRouter SDK](https://openrouter.ai/docs/sdks/python) is a Python toolkit designed to help you build AI-powered features and solutions. Giving you easy access to 400+ models across providers in an easy and type-safe way.
To learn more about how to use the OpenRouter SDK, check out our [API Reference](https://openrouter.ai/docs/sdks/python/reference) and [Documentation](https://openrouter.ai/docs/sdks/python).
@@ -199,6 +199,39 @@ with OpenRouter(
```
<!-- End Pagination [pagination] -->
<!-- Start File uploads [file-upload] -->
## File uploads
Certain SDK methods accept file objects as part of a request body or multi-part request. It is possible and typically recommended to upload files as a stream rather than reading the entire contents into memory. This avoids excessive memory consumption and potentially crashing with out-of-memory errors when working with very large files. The following example demonstrates how to attach a file stream to a request.
> [!TIP]
>
> For endpoints that handle file uploads bytes arrays can also be used. However, using streams is recommended for large files.
>
```python
from openrouter import OpenRouter
import os
with OpenRouter(
http_referer="<value>",
x_open_router_title="<value>",
x_open_router_categories="<value>",
api_key=os.getenv("OPENROUTER_API_KEY", ""),
) as open_router:
res = open_router.files.upload(file={
"file_name": "example.file",
"content": open("example.file", "rb"),
})
# Handle response
print(res)
```
<!-- End File uploads [file-upload] -->
<!-- Start Resource Management [resource-management] -->
## Resource Management
@@ -276,10 +309,4 @@ To run the test suite, you'll need to set up your environment with an OpenRouter
```bash
pytest
```
## Maturity
This SDK is in beta, and there may be breaking changes between versions without a major version update. Therefore, we recommend pinning usage
to a specific package version. This way, you can install the same version each time without breaking changes unless you are intentionally
looking for the latest version.
```
+36 -9
View File
@@ -1,6 +1,6 @@
# OpenRouter SDK (Beta)
# OpenRouter SDK
The [OpenRouter SDK](https://openrouter.ai/docs/sdks/python) is a Python toolkit designed to help you build AI-powered features and solutions. Giving you easy access to over 300 models across providers in an easy and type-safe way.
The [OpenRouter SDK](https://openrouter.ai/docs/sdks/python) is a Python toolkit designed to help you build AI-powered features and solutions. Giving you easy access to 400+ models across providers in an easy and type-safe way.
To learn more about how to use the OpenRouter SDK, check out our [API Reference](https://openrouter.ai/docs/sdks/python/reference) and [Documentation](https://openrouter.ai/docs/sdks/python).
@@ -199,6 +199,39 @@ with OpenRouter(
```
<!-- End Pagination [pagination] -->
<!-- Start File uploads [file-upload] -->
## File uploads
Certain SDK methods accept file objects as part of a request body or multi-part request. It is possible and typically recommended to upload files as a stream rather than reading the entire contents into memory. This avoids excessive memory consumption and potentially crashing with out-of-memory errors when working with very large files. The following example demonstrates how to attach a file stream to a request.
> [!TIP]
>
> For endpoints that handle file uploads bytes arrays can also be used. However, using streams is recommended for large files.
>
```python
from openrouter import OpenRouter
import os
with OpenRouter(
http_referer="<value>",
x_open_router_title="<value>",
x_open_router_categories="<value>",
api_key=os.getenv("OPENROUTER_API_KEY", ""),
) as open_router:
res = open_router.files.upload(file={
"file_name": "example.file",
"content": open("example.file", "rb"),
})
# Handle response
print(res)
```
<!-- End File uploads [file-upload] -->
<!-- Start Resource Management [resource-management] -->
## Resource Management
@@ -276,10 +309,4 @@ To run the test suite, you'll need to set up your environment with an OpenRouter
```bash
pytest
```
## Maturity
This SDK is in beta, and there may be breaking changes between versions without a major version update. Therefore, we recommend pinning usage
to a specific package version. This way, you can install the same version each time without breaking changes unless you are intentionally
looking for the latest version.
```
+81 -1
View File
@@ -18,4 +18,84 @@ Based on:
### Generated
- [python v0.9.2] .
### Releases
- [PyPI v0.9.2] https://pypi.org/project/openrouter/0.9.2 - .
- [PyPI v0.9.2] https://pypi.org/project/openrouter/0.9.2 - .
## 2026-06-17 00:59:07
### Changes
Based on:
- OpenAPI Doc
- Speakeasy CLI 1.680.0 (2.788.4) https://github.com/speakeasy-api/speakeasy
### Generated
- [python v0.10.0] .
### Releases
- [PyPI v0.10.0] https://pypi.org/project/openrouter/0.10.0 - .
## 2026-06-25 21:56:51
### Changes
Based on:
- OpenAPI Doc
- Speakeasy CLI 1.680.0 (2.788.4) https://github.com/speakeasy-api/speakeasy
### Generated
- [python v0.10.1] .
### Releases
- [PyPI v0.10.1] https://pypi.org/project/openrouter/0.10.1 - .
## 2026-06-25 22:28:38
### Changes
Based on:
- OpenAPI Doc
- Speakeasy CLI 1.680.0 (2.788.4) https://github.com/speakeasy-api/speakeasy
### Generated
- [python v0.10.2] .
### Releases
- [PyPI v0.10.2] https://pypi.org/project/openrouter/0.10.2 - .
## 2026-06-26 00:15:50
### Changes
Based on:
- OpenAPI Doc
- Speakeasy CLI 1.680.0 (2.788.4) https://github.com/speakeasy-api/speakeasy
### Generated
- [python v0.10.3] .
### Releases
- [PyPI v0.10.3] https://pypi.org/project/openrouter/0.10.3 - .
## 2026-06-26 12:06:01
### Changes
Based on:
- OpenAPI Doc
- Speakeasy CLI 1.680.0 (2.788.4) https://github.com/speakeasy-api/speakeasy
### Generated
- [python v0.10.4] .
### Releases
- [PyPI v0.10.4] https://pypi.org/project/openrouter/0.10.4 - .
## 2026-06-26 14:36:56
### Changes
Based on:
- OpenAPI Doc
- Speakeasy CLI 1.680.0 (2.788.4) https://github.com/speakeasy-api/speakeasy
### Generated
- [python v0.10.5] .
### Releases
- [PyPI v0.10.5] https://pypi.org/project/openrouter/0.10.5 - .
## 2026-06-26 18:19:11
### Changes
Based on:
- OpenAPI Doc
- Speakeasy CLI 1.680.0 (2.788.4) https://github.com/speakeasy-api/speakeasy
### Generated
- [python v0.10.6] .
### Releases
- [PyPI v0.10.6] https://pypi.org/project/openrouter/0.10.6 - .
## 2026-06-26 23:43:40
### Changes
Based on:
- OpenAPI Doc
- Speakeasy CLI 1.680.0 (2.788.4) https://github.com/speakeasy-api/speakeasy
### Generated
- [python v0.10.7] .
### Releases
- [PyPI v0.10.7] https://pypi.org/project/openrouter/0.10.7 - .
@@ -1,13 +0,0 @@
# CompletionTokensDetails
Detailed completion token usage
## Fields
| Field | Type | Required | Description |
| ---------------------------- | ---------------------------- | ---------------------------- | ---------------------------- |
| `accepted_prediction_tokens` | *OptionalNullable[int]* | :heavy_minus_sign: | Accepted prediction tokens |
| `audio_tokens` | *OptionalNullable[int]* | :heavy_minus_sign: | Tokens used for audio output |
| `reasoning_tokens` | *OptionalNullable[int]* | :heavy_minus_sign: | Tokens used for reasoning |
| `rejected_prediction_tokens` | *OptionalNullable[int]* | :heavy_minus_sign: | Rejected prediction tokens |
-11
View File
@@ -1,11 +0,0 @@
# Error
Error information
## Fields
| Field | Type | Required | Description | Example |
| ------------------- | ------------------- | ------------------- | ------------------- | ------------------- |
| `code` | *int* | :heavy_check_mark: | Error code | 429 |
| `message` | *str* | :heavy_check_mark: | Error message | Rate limit exceeded |
+21 -19
View File
@@ -5,22 +5,24 @@ Information about an AI model available on OpenRouter
## Fields
| Field | Type | Required | Description | Example |
| ------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------- |
| `architecture` | [components.ModelArchitecture](../components/modelarchitecture.md) | :heavy_check_mark: | Model architecture information | {<br/>"input_modalities": [<br/>"text"<br/>],<br/>"instruct_type": "chatml",<br/>"modality": "text-\u003etext",<br/>"output_modalities": [<br/>"text"<br/>],<br/>"tokenizer": "GPT"<br/>} |
| `canonical_slug` | *str* | :heavy_check_mark: | Canonical slug for the model | openai/gpt-4 |
| `context_length` | *Nullable[int]* | :heavy_check_mark: | Maximum context length in tokens | 8192 |
| `created` | *int* | :heavy_check_mark: | Unix timestamp of when the model was created | 1692901234 |
| `default_parameters` | [Nullable[components.DefaultParameters]](../components/defaultparameters.md) | :heavy_check_mark: | Default parameters for this model | {<br/>"frequency_penalty": 0,<br/>"presence_penalty": 0,<br/>"repetition_penalty": 1,<br/>"temperature": 0.7,<br/>"top_k": 0,<br/>"top_p": 0.9<br/>} |
| `description` | *Optional[str]* | :heavy_minus_sign: | Description of the model | GPT-4 is a large multimodal model that can solve difficult problems with greater accuracy. |
| `expiration_date` | *OptionalNullable[str]* | :heavy_minus_sign: | The date after which the model may be removed. ISO 8601 date string (YYYY-MM-DD) or null if no expiration. | 2025-06-01 |
| `hugging_face_id` | *OptionalNullable[str]* | :heavy_minus_sign: | Hugging Face model identifier, if applicable | microsoft/DialoGPT-medium |
| `id` | *str* | :heavy_check_mark: | Unique identifier for the model | openai/gpt-4 |
| `knowledge_cutoff` | *OptionalNullable[str]* | :heavy_minus_sign: | The date up to which the model was trained on data. ISO 8601 date string (YYYY-MM-DD) or null if unknown. | 2024-10-01 |
| `links` | [components.ModelLinks](../components/modellinks.md) | :heavy_check_mark: | Related API endpoints and resources for this model. | {<br/>"details": "/api/v1/models/openai/gpt-5.4/endpoints"<br/>} |
| `name` | *str* | :heavy_check_mark: | Display name of the model | GPT-4 |
| `per_request_limits` | [Nullable[components.PerRequestLimits]](../components/perrequestlimits.md) | :heavy_check_mark: | Per-request token limits | {<br/>"completion_tokens": 1000,<br/>"prompt_tokens": 1000<br/>} |
| `pricing` | [components.PublicPricing](../components/publicpricing.md) | :heavy_check_mark: | Pricing information for the model | {<br/>"completion": "0.00006",<br/>"image": "0",<br/>"prompt": "0.00003",<br/>"request": "0"<br/>} |
| `supported_parameters` | List[[components.Parameter](../components/parameter.md)] | :heavy_check_mark: | List of supported parameters for this model | |
| `supported_voices` | List[*str*] | :heavy_check_mark: | List of supported voice identifiers for TTS models. Null for non-TTS models. | <nil> |
| `top_provider` | [components.TopProviderInfo](../components/topproviderinfo.md) | :heavy_check_mark: | Information about the top provider for this model | {<br/>"context_length": 8192,<br/>"is_moderated": true,<br/>"max_completion_tokens": 4096<br/>} |
| Field | Type | Required | Description | Example |
| -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `architecture` | [components.ModelArchitecture](../components/modelarchitecture.md) | :heavy_check_mark: | Model architecture information | {<br/>"input_modalities": [<br/>"text"<br/>],<br/>"instruct_type": "chatml",<br/>"modality": "text-\u003etext",<br/>"output_modalities": [<br/>"text"<br/>],<br/>"tokenizer": "GPT"<br/>} |
| `benchmarks` | [Optional[components.ModelBenchmarks]](../components/modelbenchmarks.md) | :heavy_minus_sign: | Third-party benchmark rankings for this model. Omitted when no benchmark data is available. | {<br/>"artificial_analysis": {<br/>"agentic_index": 55.8,<br/>"coding_index": 63.2,<br/>"intelligence_index": 71.4<br/>},<br/>"design_arena": [<br/>{<br/>"arena": "models",<br/>"category": "website",<br/>"elo": 1385.2,<br/>"rank": 5,<br/>"win_rate": 62.5<br/>}<br/>]<br/>} |
| `canonical_slug` | *str* | :heavy_check_mark: | Canonical slug for the model | openai/gpt-4 |
| `context_length` | *Nullable[int]* | :heavy_check_mark: | Maximum context length in tokens | 8192 |
| `created` | *int* | :heavy_check_mark: | Unix timestamp of when the model was created | 1692901234 |
| `default_parameters` | [Nullable[components.DefaultParameters]](../components/defaultparameters.md) | :heavy_check_mark: | Default parameters for this model | {<br/>"frequency_penalty": 0,<br/>"presence_penalty": 0,<br/>"repetition_penalty": 1,<br/>"temperature": 0.7,<br/>"top_k": 0,<br/>"top_p": 0.9<br/>} |
| `description` | *Optional[str]* | :heavy_minus_sign: | Description of the model | GPT-4 is a large multimodal model that can solve difficult problems with greater accuracy. |
| `expiration_date` | *OptionalNullable[str]* | :heavy_minus_sign: | The date after which the model may be removed. ISO 8601 date string (YYYY-MM-DD) or null if no expiration. | 2025-06-01 |
| `hugging_face_id` | *OptionalNullable[str]* | :heavy_minus_sign: | Hugging Face model identifier, if applicable | microsoft/DialoGPT-medium |
| `id` | *str* | :heavy_check_mark: | Unique identifier for the model | openai/gpt-4 |
| `knowledge_cutoff` | *OptionalNullable[str]* | :heavy_minus_sign: | The date up to which the model was trained on data. ISO 8601 date string (YYYY-MM-DD) or null if unknown. | 2024-10-01 |
| `links` | [components.ModelLinks](../components/modellinks.md) | :heavy_check_mark: | Related API endpoints and resources for this model. | {<br/>"details": "/api/v1/models/openai/gpt-5.4/endpoints"<br/>} |
| `name` | *str* | :heavy_check_mark: | Display name of the model | GPT-4 |
| `per_request_limits` | [Nullable[components.PerRequestLimits]](../components/perrequestlimits.md) | :heavy_check_mark: | Per-request token limits | {<br/>"completion_tokens": 1000,<br/>"prompt_tokens": 1000<br/>} |
| `pricing` | [components.PublicPricing](../components/publicpricing.md) | :heavy_check_mark: | Pricing information for the model | {<br/>"completion": "0.00006",<br/>"image": "0",<br/>"prompt": "0.00003",<br/>"request": "0"<br/>} |
| `reasoning` | [Optional[components.ModelReasoning]](../components/modelreasoning.md) | :heavy_minus_sign: | Reasoning effort configuration. Omitted for non-reasoning models and dynamic router models. | {<br/>"default_effort": "medium",<br/>"default_enabled": true,<br/>"mandatory": false,<br/>"supported_efforts": [<br/>"high",<br/>"medium",<br/>"low",<br/>"minimal"<br/>]<br/>} |
| `supported_parameters` | List[[components.Parameter](../components/parameter.md)] | :heavy_check_mark: | List of supported parameters for this model | |
| `supported_voices` | List[*str*] | :heavy_check_mark: | List of supported voice identifiers for TTS models. Null for non-TTS models. | <nil> |
| `top_provider` | [components.TopProviderInfo](../components/topproviderinfo.md) | :heavy_check_mark: | Information about the top provider for this model | {<br/>"context_length": 8192,<br/>"is_moderated": true,<br/>"max_completion_tokens": 4096<br/>} |
+17 -16
View File
@@ -3,19 +3,20 @@
## Fields
| Field | Type | Required | Description | Example |
| -------------------- | -------------------- | -------------------- | -------------------- | -------------------- |
| `audio` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `audio_output` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `completion` | *str* | :heavy_check_mark: | N/A | 1000 |
| `discount` | *Optional[float]* | :heavy_minus_sign: | N/A | |
| `image` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `image_output` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `image_token` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `input_audio_cache` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `input_cache_read` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `input_cache_write` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `internal_reasoning` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `prompt` | *str* | :heavy_check_mark: | N/A | 1000 |
| `request` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `web_search` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| Field | Type | Required | Description |
| --------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `audio` | *Optional[str]* | :heavy_minus_sign: | Price in USD per audio input token |
| `audio_output` | *Optional[str]* | :heavy_minus_sign: | Price in USD per audio output token |
| `completion` | *str* | :heavy_check_mark: | Price in USD per token for completion (output) generation |
| `discount` | *Optional[float]* | :heavy_minus_sign: | Fractional discount applied to this endpoint's pricing; the price is multiplied by (1 - discount) (0 = no discount, 1 = free) |
| `image` | *Optional[str]* | :heavy_minus_sign: | Price in USD per input image |
| `image_output` | *Optional[str]* | :heavy_minus_sign: | Price in USD per output image |
| `image_token` | *Optional[str]* | :heavy_minus_sign: | Price in USD per image token |
| `input_audio_cache` | *Optional[str]* | :heavy_minus_sign: | Price in USD per cached audio input token |
| `input_cache_read` | *Optional[str]* | :heavy_minus_sign: | Price in USD per cached input token (read) |
| `input_cache_write` | *Optional[str]* | :heavy_minus_sign: | Price per cache-write token, in USD per token. For providers with multiple cache TTLs (e.g. Anthropic), this is the default (5-minute) cache-write rate. |
| `input_cache_write_1h` | *Optional[str]* | :heavy_minus_sign: | Price per 1-hour cache-write token, in USD per token. Only present for providers that price an extended (1-hour) cache TTL separately, such as Anthropic. |
| `internal_reasoning` | *Optional[str]* | :heavy_minus_sign: | Price in USD per internal reasoning token |
| `prompt` | *str* | :heavy_check_mark: | Price in USD per token for prompt (input) processing |
| `request` | *Optional[str]* | :heavy_minus_sign: | Price in USD per request |
| `web_search` | *Optional[str]* | :heavy_minus_sign: | Price in USD per web search |
-13
View File
@@ -1,13 +0,0 @@
# PromptTokensDetails
Detailed prompt token usage
## Fields
| Field | Type | Required | Description |
| ------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------ |
| `audio_tokens` | *Optional[int]* | :heavy_minus_sign: | Audio input tokens |
| `cache_write_tokens` | *Optional[int]* | :heavy_minus_sign: | Tokens written to cache. Only returned for models with explicit caching and cache write pricing. |
| `cached_tokens` | *Optional[int]* | :heavy_minus_sign: | Cached prompt tokens |
| `video_tokens` | *Optional[int]* | :heavy_minus_sign: | Video input tokens |
+5
View File
@@ -42,12 +42,14 @@
| `GOOGLE` | Google |
| `GOOGLE_AI_STUDIO` | Google AI Studio |
| `GROQ` | Groq |
| `HEY_GEN` | HeyGen |
| `INCEPTION` | Inception |
| `INCEPTRON` | Inceptron |
| `INFERENCE_NET` | InferenceNet |
| `IONSTREAM` | Ionstream |
| `INFERMATIC` | Infermatic |
| `IO_NET` | Io Net |
| `INFERACT_V_LLM` | Inferact vLLM |
| `INFLECTION` | Inflection |
| `LIQUID` | Liquid |
| `MARA` | Mara |
@@ -74,6 +76,7 @@
| `RECRAFT` | Recraft |
| `REKA` | Reka |
| `RELACE` | Relace |
| `SAKANA_AI` | Sakana AI |
| `SAMBA_NOVA` | SambaNova |
| `SEED` | Seed |
| `SILICON_FLOW` | SiliconFlow |
@@ -82,11 +85,13 @@
| `STEALTH` | Stealth |
| `STREAM_LAKE` | StreamLake |
| `SWITCHPOINT` | Switchpoint |
| `TENSTORRENT` | Tenstorrent |
| `TOGETHER` | Together |
| `UPSTAGE` | Upstage |
| `VENICE` | Venice |
| `WAFER` | Wafer |
| `WAND_B` | WandB |
| `QUIVER` | Quiver |
| `XIAOMI` | Xiaomi |
| `X_AI` | xAI |
| `Z_AI` | Z.AI |
+17 -16
View File
@@ -5,19 +5,20 @@ Pricing information for the model
## Fields
| Field | Type | Required | Description | Example |
| -------------------- | -------------------- | -------------------- | -------------------- | -------------------- |
| `audio` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `audio_output` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `completion` | *str* | :heavy_check_mark: | N/A | 1000 |
| `discount` | *Optional[float]* | :heavy_minus_sign: | N/A | |
| `image` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `image_output` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `image_token` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `input_audio_cache` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `input_cache_read` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `input_cache_write` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `internal_reasoning` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `prompt` | *str* | :heavy_check_mark: | N/A | 1000 |
| `request` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| `web_search` | *Optional[str]* | :heavy_minus_sign: | N/A | 1000 |
| Field | Type | Required | Description |
| --------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `audio` | *Optional[str]* | :heavy_minus_sign: | Price in USD per audio input token |
| `audio_output` | *Optional[str]* | :heavy_minus_sign: | Price in USD per audio output token |
| `completion` | *str* | :heavy_check_mark: | Price in USD per token for completion (output) generation |
| `discount` | *Optional[float]* | :heavy_minus_sign: | Fractional discount applied to this endpoint's pricing; the price is multiplied by (1 - discount) (0 = no discount, 1 = free) |
| `image` | *Optional[str]* | :heavy_minus_sign: | Price in USD per input image |
| `image_output` | *Optional[str]* | :heavy_minus_sign: | Price in USD per output image |
| `image_token` | *Optional[str]* | :heavy_minus_sign: | Price in USD per image token |
| `input_audio_cache` | *Optional[str]* | :heavy_minus_sign: | Price in USD per cached audio input token |
| `input_cache_read` | *Optional[str]* | :heavy_minus_sign: | Price in USD per cached input token (read) |
| `input_cache_write` | *Optional[str]* | :heavy_minus_sign: | Price per cache-write token, in USD per token. For providers with multiple cache TTLs (e.g. Anthropic), this is the default (5-minute) cache-write rate. |
| `input_cache_write_1h` | *Optional[str]* | :heavy_minus_sign: | Price per 1-hour cache-write token, in USD per token. Only present for providers that price an extended (1-hour) cache TTL separately, such as Anthropic. |
| `internal_reasoning` | *Optional[str]* | :heavy_minus_sign: | Price in USD per internal reasoning token |
| `prompt` | *str* | :heavy_check_mark: | Price in USD per token for prompt (input) processing |
| `request` | *Optional[str]* | :heavy_minus_sign: | Price in USD per request |
| `web_search` | *Optional[str]* | :heavy_minus_sign: | Price in USD per web search |
+1
View File
@@ -5,6 +5,7 @@
| Field | Type | Required | Description |
| -------------------------------------------------------------- | -------------------------------------------------------------- | -------------------------------------------------------------- | -------------------------------------------------------------- |
| `content` | *Optional[str]* | :heavy_minus_sign: | N/A |
| `end_index` | *int* | :heavy_check_mark: | N/A |
| `start_index` | *int* | :heavy_check_mark: | N/A |
| `title` | *str* | :heavy_check_mark: | N/A |
+7 -6
View File
@@ -5,9 +5,10 @@ The search engine to use for web search.
## Values
| Name | Value |
| ----------- | ----------- |
| `NATIVE` | native |
| `EXA` | exa |
| `FIRECRAWL` | firecrawl |
| `PARALLEL` | parallel |
| Name | Value |
| ------------ | ------------ |
| `NATIVE` | native |
| `EXA` | exa |
| `FIRECRAWL` | firecrawl |
| `PARALLEL` | parallel |
| `PERPLEXITY` | perplexity |
@@ -3,12 +3,13 @@
## Fields
| Field | Type | Required | Description | Example |
| -------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------- |
| `callback_url` | *str* | :heavy_check_mark: | The callback URL to redirect to after authorization. Note, only https URLs on ports 443 and 3000 are allowed. | https://myapp.com/auth/callback |
| `code_challenge` | *Optional[str]* | :heavy_minus_sign: | PKCE code challenge for enhanced security | E9Melhoa2OwvFrEMTJguCHaoeK1t8URWbuGJSstw-cM |
| `code_challenge_method` | [Optional[operations.CreateAuthKeysCodeCodeChallengeMethod]](../operations/createauthkeyscodecodechallengemethod.md) | :heavy_minus_sign: | The method used to generate the code challenge | S256 |
| `expires_at` | [date](https://docs.python.org/3/library/datetime.html#date-objects) | :heavy_minus_sign: | Optional expiration time for the API key to be created | 2027-12-31T23:59:59Z |
| `key_label` | *Optional[str]* | :heavy_minus_sign: | Optional custom label for the API key. Defaults to the app name if not provided. | My Custom Key |
| `limit` | *Optional[float]* | :heavy_minus_sign: | Credit limit for the API key to be created | 100 |
| `usage_limit_type` | [Optional[operations.UsageLimitType]](../operations/usagelimittype.md) | :heavy_minus_sign: | Optional credit limit reset interval. When set, the credit limit resets on this interval. | monthly |
| Field | Type | Required | Description | Example |
| -------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------- |
| `callback_url` | *str* | :heavy_check_mark: | The callback URL to redirect to after authorization. Supports https URLs and localhost/127.0.0.1 URLs on any port for local CLI tools. | https://myapp.com/auth/callback |
| `code_challenge` | *Optional[str]* | :heavy_minus_sign: | PKCE code challenge for enhanced security | E9Melhoa2OwvFrEMTJguCHaoeK1t8URWbuGJSstw-cM |
| `code_challenge_method` | [Optional[operations.CreateAuthKeysCodeCodeChallengeMethod]](../operations/createauthkeyscodecodechallengemethod.md) | :heavy_minus_sign: | The method used to generate the code challenge | S256 |
| `expires_at` | [date](https://docs.python.org/3/library/datetime.html#date-objects) | :heavy_minus_sign: | Optional expiration time for the API key to be created | 2027-12-31T23:59:59Z |
| `key_label` | *Optional[str]* | :heavy_minus_sign: | Optional custom label for the API key. Defaults to the app name if not provided. | My Custom Key |
| `limit` | *Optional[float]* | :heavy_minus_sign: | Credit limit for the API key to be created | 100 |
| `usage_limit_type` | [Optional[operations.UsageLimitType]](../operations/usagelimittype.md) | :heavy_minus_sign: | Optional credit limit reset interval. When set, the credit limit resets on this interval. | monthly |
| `workspace_id` | *Optional[str]* | :heavy_minus_sign: | Optional workspace ID to associate the API key with | |
+20 -9
View File
@@ -3,12 +3,23 @@
## Fields
| Field | Type | Required | Description | Example |
| ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `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_categories` | *Optional[str]* | :heavy_minus_sign: | Comma-separated list of app categories (e.g. "cli-agent,cloud-agent"). Used for marketplace rankings.<br/> | |
| `category` | [Optional[operations.GetModelsCategory]](../operations/getmodelscategory.md) | :heavy_minus_sign: | Filter models by use case category | programming |
| `supported_parameters` | *Optional[str]* | :heavy_minus_sign: | Filter models by supported parameter (comma-separated) | temperature |
| `output_modalities` | *Optional[str]* | :heavy_minus_sign: | Filter models by output modality. Accepts a comma-separated list of modalities (text, image, audio, embeddings) or "all" to include all models. Defaults to "text". | text |
| `sort` | [Optional[operations.GetModelsSort]](../operations/getmodelssort.md) | :heavy_minus_sign: | Sort the returned models server-side. Prefer this over fetching the full list and sorting client-side. Options: pricing-low-to-high, pricing-high-to-low (average prompt/completion price), context-high-to-low (context length), throughput-high-to-low, latency-low-to-high (recent median performance), most-popular, top-weekly (tokens processed in the last week), newest (creation date). When omitted, the existing default ordering is preserved. | newest |
| Field | Type | Required | Description | Example |
| ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `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_categories` | *Optional[str]* | :heavy_minus_sign: | Comma-separated list of app categories (e.g. "cli-agent,cloud-agent"). Used for marketplace rankings.<br/> | |
| `category` | [Optional[operations.GetModelsCategory]](../operations/getmodelscategory.md) | :heavy_minus_sign: | Filter models by use case category | programming |
| `supported_parameters` | *Optional[str]* | :heavy_minus_sign: | Filter models by supported parameter (comma-separated) | temperature |
| `output_modalities` | *Optional[str]* | :heavy_minus_sign: | Filter models by output modality. Accepts a comma-separated list of modalities (text, image, audio, embeddings) or "all" to include all models. Defaults to "text". | text |
| `sort` | [Optional[operations.GetModelsSort]](../operations/getmodelssort.md) | :heavy_minus_sign: | Sort the returned models server-side. Prefer this over fetching the full list and sorting client-side. Options: pricing-low-to-high, pricing-high-to-low (average prompt/completion price), context-high-to-low (context length), throughput-high-to-low, latency-low-to-high (recent median performance), most-popular, top-weekly (tokens processed in the last week), newest (creation date), intelligence-high-to-low (Artificial Analysis intelligence index), design-arena-elo-high-to-low (best Design Arena ELO across arenas). Models without a score for the chosen benchmark are placed last. When omitted, the existing default ordering is preserved. | newest |
| `q` | *Optional[str]* | :heavy_minus_sign: | Free-text search by model name or slug. | gpt-4 |
| `input_modalities` | *Optional[str]* | :heavy_minus_sign: | Filter models by input modality. Comma-separated list of: text, image, audio, file. | text,image |
| `context` | *Optional[int]* | :heavy_minus_sign: | Minimum context length (tokens). Models with smaller context are excluded. | 128000 |
| `min_price` | *OptionalNullable[float]* | :heavy_minus_sign: | Minimum prompt price in $/M tokens. | 0 |
| `max_price` | *OptionalNullable[float]* | :heavy_minus_sign: | Maximum prompt price in $/M tokens. | 10 |
| `arch` | *Optional[str]* | :heavy_minus_sign: | Filter models by architecture/model family (e.g. GPT, Claude, Gemini, Llama). | GPT |
| `model_authors` | *Optional[str]* | :heavy_minus_sign: | Filter models by the organization that created the model. Comma-separated list of author slugs. | openai,anthropic |
| `providers` | *Optional[str]* | :heavy_minus_sign: | Filter models by hosting provider. Comma-separated list of provider names. | OpenAI,Anthropic |
| `distillable` | [Optional[operations.Distillable]](../operations/distillable.md) | :heavy_minus_sign: | Filter by distillation capability. "true" returns only distillable models, "false" excludes them. | true |
| `zdr` | [Optional[operations.Zdr]](../operations/zdr.md) | :heavy_minus_sign: | When set to "true", return only models with zero data retention endpoints. | true |
| `region` | [Optional[operations.Region]](../operations/region.md) | :heavy_minus_sign: | Filter to models with endpoints in the given data region. Currently only "eu" is supported. | eu |
+1 -1
View File
@@ -113,7 +113,7 @@ with OpenRouter(
## create
Create a new API key for the authenticated user. [Management key](/docs/guides/overview/auth/management-api-keys) required.
Create a new API key for the authenticated user. The plaintext `key` is returned only in this response. Treat it as a write-only, sensitive value; it cannot be retrieved later. [Management key](/docs/guides/overview/auth/management-api-keys) required.
### Example Usage
+5
View File
@@ -61,6 +61,7 @@ with OpenRouter(
| `max_completion_tokens` | *OptionalNullable[int]* | :heavy_minus_sign: | Maximum tokens in completion | 100 |
| `max_tokens` | *OptionalNullable[int]* | :heavy_minus_sign: | Maximum tokens (deprecated, use max_completion_tokens). Note: some providers enforce a minimum of 16. | 100 |
| `metadata` | Dict[str, *str*] | :heavy_minus_sign: | Key-value pairs for additional object information (max 16 pairs, 64 char keys, 512 char values) | {<br/>"session_id": "session-456",<br/>"user_id": "user-123"<br/>} |
| `min_p` | *OptionalNullable[float]* | :heavy_minus_sign: | Minimum probability threshold relative to the most likely token. Tokens with probability below min_p * (probability of top token) are filtered out. Not all providers support this parameter. | 0.1 |
| `modalities` | List[[components.Modality](../../components/modality.md)] | :heavy_minus_sign: | Output modalities for the response. Supported values are "text", "image", and "audio". | [<br/>"text",<br/>"image"<br/>] |
| `model` | *Optional[str]* | :heavy_minus_sign: | Model to use for completion | openai/gpt-4 |
| `models` | List[*str*] | :heavy_minus_sign: | Models to use for completion | [<br/>"openai/gpt-4",<br/>"openai/gpt-4o"<br/>] |
@@ -69,6 +70,8 @@ with OpenRouter(
| `presence_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | Presence penalty (-2.0 to 2.0) | 0 |
| `provider` | [OptionalNullable[components.ProviderPreferences]](../../components/providerpreferences.md) | :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.md) | :heavy_minus_sign: | Configuration options for reasoning models | {<br/>"effort": "medium",<br/>"summary": "concise"<br/>} |
| `reasoning_effort` | [OptionalNullable[components.ChatRequestReasoningEffort]](../../components/chatrequestreasoningeffort.md) | :heavy_minus_sign: | Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ. | medium |
| `repetition_penalty` | *OptionalNullable[float]* | :heavy_minus_sign: | Penalizes tokens based on how much they have already appeared in the text. A value of 1.0 means no penalty. Values above 1.0 penalize repeated tokens more strongly. Not all providers support this parameter. | 1 |
| `response_format` | [Optional[components.ResponseFormat]](../../components/responseformat.md) | :heavy_minus_sign: | Response format configuration | {<br/>"type": "json_object"<br/>} |
| `seed` | *OptionalNullable[int]* | :heavy_minus_sign: | Random seed for deterministic outputs | 42 |
| `service_tier` | [OptionalNullable[components.ChatRequestServiceTier]](../../components/chatrequestservicetier.md) | :heavy_minus_sign: | The service tier to use for processing this request. | auto |
@@ -80,6 +83,8 @@ with OpenRouter(
| `temperature` | *OptionalNullable[float]* | :heavy_minus_sign: | Sampling temperature (0-2) | 0.7 |
| `tool_choice` | [Optional[components.ChatToolChoice]](../../components/chattoolchoice.md) | :heavy_minus_sign: | Tool choice configuration | auto |
| `tools` | List[[components.ChatFunctionTool](../../components/chatfunctiontool.md)] | :heavy_minus_sign: | Available tools for function calling | [<br/>{<br/>"function": {<br/>"description": "Get weather",<br/>"name": "get_weather"<br/>},<br/>"type": "function"<br/>}<br/>] |
| `top_a` | *OptionalNullable[float]* | :heavy_minus_sign: | Consider only tokens with "sufficiently high" probabilities based on the probability of the most likely token. Not all providers support this parameter. | 0 |
| `top_k` | *OptionalNullable[int]* | :heavy_minus_sign: | Limits the model to choose from the top K most likely tokens at each step. A value of 1 means the model will always pick the most likely next token. Not all providers support this parameter. | 40 |
| `top_logprobs` | *OptionalNullable[int]* | :heavy_minus_sign: | Number of top log probabilities to return (0-20) | 5 |
| `top_p` | *OptionalNullable[float]* | :heavy_minus_sign: | Nucleus sampling parameter (0-1) | 1 |
| `trace` | [Optional[components.TraceConfig]](../../components/traceconfig.md) | :heavy_minus_sign: | Metadata for observability and tracing. Known keys (trace_id, trace_name, span_name, generation_name, parent_span_id) have special handling. Additional keys are passed through as custom metadata to configured broadcast destinations. | {<br/>"trace_id": "trace-abc123",<br/>"trace_name": "my-app-trace"<br/>} |
+2 -2
View File
@@ -241,7 +241,7 @@ with OpenRouter(
## update
Update an existing guardrail. [Management key](/docs/guides/overview/auth/management-api-keys) required.
Update an existing guardrail. Collection fields use replace semantics: send the full desired set on every update. [Management key](/docs/guides/overview/auth/management-api-keys) required.
### Example Usage
@@ -359,7 +359,7 @@ with OpenRouter(
## bulk_assign_keys
Assign multiple API keys to a specific guardrail. [Management key](/docs/guides/overview/auth/management-api-keys) required.
Assign multiple API keys to a specific guardrail. A key may hold at most one guardrail; assigning replaces any existing assignment. [Management key](/docs/guides/overview/auth/management-api-keys) required.
### Example Usage
+71 -10
View File
@@ -6,10 +6,60 @@ Model information endpoints
### Available Operations
* [get](#get) - Get a model by its slug
* [list](#list) - List all models and their properties
* [count](#count) - Get total count of available models
* [list_for_user](#list_for_user) - List models filtered by user provider preferences, privacy settings, and guardrails
## get
Returns full details for a single model identified by its author and slug (e.g. openai/gpt-4). Supports variant suffixes (e.g. openai/gpt-4:free) and resolves known slug aliases.
### Example Usage
<!-- UsageSnippet language="python" operationID="getModel" method="get" path="/model/{author}/{slug}" -->
```python
from openrouter import OpenRouter
import os
with OpenRouter(
http_referer="<value>",
x_open_router_title="<value>",
x_open_router_categories="<value>",
api_key=os.getenv("OPENROUTER_API_KEY", ""),
) as open_router:
res = open_router.models.get(author="openai", slug="gpt-4")
# Handle response
print(res)
```
### Parameters
| Parameter | Type | Required | Description | Example |
| ------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------- |
| `author` | *str* | :heavy_check_mark: | The author/organization of the model | openai |
| `slug` | *str* | :heavy_check_mark: | The model slug, optionally including a variant suffix (e.g. gpt-4 or gpt-4:free) | gpt-4 |
| `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_categories` | *Optional[str]* | :heavy_minus_sign: | Comma-separated list of app categories (e.g. "cli-agent,cloud-agent"). Used for marketplace rankings.<br/> | |
| `retries` | [Optional[utils.RetryConfig]](../../models/utils/retryconfig.md) | :heavy_minus_sign: | Configuration to override the default retry behavior of the client. | |
### Response
**[components.ModelResponse](../../components/modelresponse.md)**
### Errors
| Error Type | Status Code | Content Type |
| ---------------------------------- | ---------------------------------- | ---------------------------------- |
| errors.NotFoundResponseError | 404 | application/json |
| errors.InternalServerResponseError | 500 | application/json |
| errors.OpenRouterDefaultError | 4XX, 5XX | \*/\* |
## list
List all models and their properties
@@ -38,16 +88,27 @@ with OpenRouter(
### Parameters
| Parameter | Type | Required | Description | Example |
| ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `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_categories` | *Optional[str]* | :heavy_minus_sign: | Comma-separated list of app categories (e.g. "cli-agent,cloud-agent"). Used for marketplace rankings.<br/> | |
| `category` | [Optional[operations.GetModelsCategory]](../../operations/getmodelscategory.md) | :heavy_minus_sign: | Filter models by use case category | programming |
| `supported_parameters` | *Optional[str]* | :heavy_minus_sign: | Filter models by supported parameter (comma-separated) | temperature |
| `output_modalities` | *Optional[str]* | :heavy_minus_sign: | Filter models by output modality. Accepts a comma-separated list of modalities (text, image, audio, embeddings) or "all" to include all models. Defaults to "text". | text |
| `sort` | [Optional[operations.GetModelsSort]](../../operations/getmodelssort.md) | :heavy_minus_sign: | Sort the returned models server-side. Prefer this over fetching the full list and sorting client-side. Options: pricing-low-to-high, pricing-high-to-low (average prompt/completion price), context-high-to-low (context length), throughput-high-to-low, latency-low-to-high (recent median performance), most-popular, top-weekly (tokens processed in the last week), newest (creation date). When omitted, the existing default ordering is preserved. | newest |
| `retries` | [Optional[utils.RetryConfig]](../../models/utils/retryconfig.md) | :heavy_minus_sign: | Configuration to override the default retry behavior of the client. | |
| Parameter | Type | Required | Description | Example |
| ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `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_categories` | *Optional[str]* | :heavy_minus_sign: | Comma-separated list of app categories (e.g. "cli-agent,cloud-agent"). Used for marketplace rankings.<br/> | |
| `category` | [Optional[operations.GetModelsCategory]](../../operations/getmodelscategory.md) | :heavy_minus_sign: | Filter models by use case category | programming |
| `supported_parameters` | *Optional[str]* | :heavy_minus_sign: | Filter models by supported parameter (comma-separated) | temperature |
| `output_modalities` | *Optional[str]* | :heavy_minus_sign: | Filter models by output modality. Accepts a comma-separated list of modalities (text, image, audio, embeddings) or "all" to include all models. Defaults to "text". | text |
| `sort` | [Optional[operations.GetModelsSort]](../../operations/getmodelssort.md) | :heavy_minus_sign: | Sort the returned models server-side. Prefer this over fetching the full list and sorting client-side. Options: pricing-low-to-high, pricing-high-to-low (average prompt/completion price), context-high-to-low (context length), throughput-high-to-low, latency-low-to-high (recent median performance), most-popular, top-weekly (tokens processed in the last week), newest (creation date), intelligence-high-to-low (Artificial Analysis intelligence index), design-arena-elo-high-to-low (best Design Arena ELO across arenas). Models without a score for the chosen benchmark are placed last. When omitted, the existing default ordering is preserved. | newest |
| `q` | *Optional[str]* | :heavy_minus_sign: | Free-text search by model name or slug. | gpt-4 |
| `input_modalities` | *Optional[str]* | :heavy_minus_sign: | Filter models by input modality. Comma-separated list of: text, image, audio, file. | text,image |
| `context` | *Optional[int]* | :heavy_minus_sign: | Minimum context length (tokens). Models with smaller context are excluded. | 128000 |
| `min_price` | *OptionalNullable[float]* | :heavy_minus_sign: | Minimum prompt price in $/M tokens. | 0 |
| `max_price` | *OptionalNullable[float]* | :heavy_minus_sign: | Maximum prompt price in $/M tokens. | 10 |
| `arch` | *Optional[str]* | :heavy_minus_sign: | Filter models by architecture/model family (e.g. GPT, Claude, Gemini, Llama). | GPT |
| `model_authors` | *Optional[str]* | :heavy_minus_sign: | Filter models by the organization that created the model. Comma-separated list of author slugs. | openai,anthropic |
| `providers` | *Optional[str]* | :heavy_minus_sign: | Filter models by hosting provider. Comma-separated list of provider names. | OpenAI,Anthropic |
| `distillable` | [Optional[operations.Distillable]](../../operations/distillable.md) | :heavy_minus_sign: | Filter by distillation capability. "true" returns only distillable models, "false" excludes them. | true |
| `zdr` | [Optional[operations.Zdr]](../../operations/zdr.md) | :heavy_minus_sign: | When set to "true", return only models with zero data retention endpoints. | true |
| `region` | [Optional[operations.Region]](../../operations/region.md) | :heavy_minus_sign: | Filter to models with endpoints in the given data region. Currently only "eu" is supported. | eu |
| `retries` | [Optional[utils.RetryConfig]](../../models/utils/retryconfig.md) | :heavy_minus_sign: | Configuration to override the default retry behavior of the client. | |
### Response
+3 -1
View File
@@ -90,7 +90,7 @@ with OpenRouter(
| Parameter | Type | Required | Description | Example |
| ------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------- |
| `callback_url` | *str* | :heavy_check_mark: | The callback URL to redirect to after authorization. Note, only https URLs on ports 443 and 3000 are allowed. | https://myapp.com/auth/callback |
| `callback_url` | *str* | :heavy_check_mark: | The callback URL to redirect to after authorization. Supports https URLs and localhost/127.0.0.1 URLs on any port for local CLI tools. | https://myapp.com/auth/callback |
| `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_categories` | *Optional[str]* | :heavy_minus_sign: | Comma-separated list of app categories (e.g. "cli-agent,cloud-agent"). Used for marketplace rankings.<br/> | |
@@ -100,6 +100,7 @@ with OpenRouter(
| `key_label` | *Optional[str]* | :heavy_minus_sign: | Optional custom label for the API key. Defaults to the app name if not provided. | My Custom Key |
| `limit` | *Optional[float]* | :heavy_minus_sign: | Credit limit for the API key to be created | 100 |
| `usage_limit_type` | [Optional[operations.UsageLimitType]](../../operations/usagelimittype.md) | :heavy_minus_sign: | Optional credit limit reset interval. When set, the credit limit resets on this interval. | monthly |
| `workspace_id` | *Optional[str]* | :heavy_minus_sign: | Optional workspace ID to associate the API key with | |
| `retries` | [Optional[utils.RetryConfig]](../../models/utils/retryconfig.md) | :heavy_minus_sign: | Configuration to override the default retry behavior of the client. | |
### Response
@@ -112,6 +113,7 @@ with OpenRouter(
| ---------------------------------- | ---------------------------------- | ---------------------------------- |
| errors.BadRequestResponseError | 400 | application/json |
| errors.UnauthorizedResponseError | 401 | application/json |
| errors.ForbiddenResponseError | 403 | application/json |
| errors.ConflictResponseError | 409 | application/json |
| errors.InternalServerResponseError | 500 | application/json |
| errors.OpenRouterDefaultError | 4XX, 5XX | \*/\* |
+1
View File
@@ -46,6 +46,7 @@ with OpenRouter(
| `x_open_router_metadata` | [Optional[components.MetadataLevel]](../../components/metadatalevel.md) | :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 | |
| `cache_control` | [Optional[components.AnthropicCacheControlDirective]](../../components/anthropiccachecontroldirective.md) | :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/>} |
| `debug` | [Optional[components.ChatDebugOptions]](../../components/chatdebugoptions.md) | :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 | |
| `image_config` | Dict[str, [components.ImageConfig](../../components/imageconfig.md)] | :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/>} |
| `include` | List[[components.ResponseIncludesEnum](../../components/responseincludesenum.md)] | :heavy_minus_sign: | N/A | |
+1 -1
View File
@@ -1,6 +1,6 @@
[project]
name = "openrouter"
version = "0.9.2"
version = "0.10.7"
description = "Official Python Client SDK for OpenRouter."
authors = [{ name = "OpenRouter" },]
readme = "README-PYPI.md"
+2 -2
View File
@@ -3,10 +3,10 @@
import importlib.metadata
__title__: str = "openrouter"
__version__: str = "0.9.2"
__version__: str = "0.10.7"
__openapi_doc_version__: str = "1.0.0"
__gen_version__: str = "2.788.4"
__user_agent__: str = "speakeasy-sdk/python 0.9.2 2.788.4 1.0.0 openrouter"
__user_agent__: str = "speakeasy-sdk/python 0.10.7 2.788.4 1.0.0 openrouter"
try:
if __package__ is not None:
+2 -2
View File
@@ -533,7 +533,7 @@ class APIKeys(BaseSDK):
) -> operations.CreateKeysResponse:
r"""Create a new API key
Create a new API key for the authenticated user. [Management key](/docs/guides/overview/auth/management-api-keys) required.
Create a new API key for the authenticated user. The plaintext `key` is returned only in this response. Treat it as a write-only, sensitive value; it cannot be retrieved later. [Management key](/docs/guides/overview/auth/management-api-keys) required.
:param name: Name for the new API key
:param http_referer: The app identifier should be your app's URL and is used as the primary identifier for rankings.
@@ -696,7 +696,7 @@ class APIKeys(BaseSDK):
) -> operations.CreateKeysResponse:
r"""Create a new API key
Create a new API key for the authenticated user. [Management key](/docs/guides/overview/auth/management-api-keys) required.
Create a new API key for the authenticated user. The plaintext `key` is returned only in this response. Treat it as a write-only, sensitive value; it cannot be retrieved later. [Management key](/docs/guides/overview/auth/management-api-keys) required.
:param name: Name for the new API key
:param http_referer: The app identifier should be your app's URL and is used as the primary identifier for rankings.
+303
View File
@@ -0,0 +1,303 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from .basesdk import BaseSDK
from openrouter import components, errors, operations, utils
from openrouter._hooks import HookContext
from openrouter.types import OptionalNullable, UNSET
from openrouter.utils import get_security_from_env
from openrouter.utils.unmarshal_json_response import unmarshal_json_response
from typing import Any, Mapping, Optional
class Benchmarks(BaseSDK):
r"""Benchmarks endpoints"""
def get_benchmarks(
self,
*,
http_referer: Optional[str] = None,
x_open_router_title: Optional[str] = None,
x_open_router_categories: Optional[str] = None,
source: Optional[operations.Source] = None,
task_type: Optional[operations.TaskType] = None,
arena: Optional[operations.Arena] = None,
category: Optional[str] = None,
max_results: Optional[int] = None,
retries: OptionalNullable[utils.RetryConfig] = UNSET,
server_url: Optional[str] = None,
timeout_ms: Optional[int] = None,
http_headers: Optional[Mapping[str, str]] = None,
) -> components.UnifiedBenchmarksResponse:
r"""List Benchmarks
Unified benchmark endpoint that aggregates scores from multiple benchmark sources (Artificial Analysis, Design Arena). Filter by source to reproduce the exact shapes from the legacy per-source endpoints, or use task_type to find models suited for specific workloads. Authenticate with any valid OpenRouter API key. Rate-limited to 30 requests/minute per key and 500 requests/day per account.
:param http_referer: The app identifier should be your app's URL and is used as the primary identifier for rankings.
This is used to track API usage per application.
:param x_open_router_title: The app display name allows you to customize how your app appears in OpenRouter's dashboard.
:param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings.
:param source: Benchmark source to query. Determines the shape of the returned items. When omitted, returns results from all sources.
:param task_type: Filter results by task type. For Artificial Analysis, maps to the corresponding index. For Design Arena, maps to the matching category.
:param arena: Design Arena only: arena to query. Defaults to `models` when source is `design-arena`.
:param category: Design Arena only: category within the arena (e.g. `codecategories`, `uicomponent`, `gamedev`, `3d`, `dataviz`, `image`, `video`, `svg`). When omitted, returns all categories.
:param max_results: Maximum number of items to return. When omitted, all matching results are returned.
:param retries: Override the default retry configuration for this method
:param server_url: Override the default server URL for this method
:param timeout_ms: Override the default request timeout configuration for this method in milliseconds
:param http_headers: Additional headers to set or replace on requests.
"""
base_url = None
url_variables = None
if timeout_ms is None:
timeout_ms = self.sdk_configuration.timeout_ms
if server_url is not None:
base_url = server_url
else:
base_url = self._get_url(base_url, url_variables)
request = operations.GetBenchmarksRequest(
http_referer=http_referer,
x_open_router_title=x_open_router_title,
x_open_router_categories=x_open_router_categories,
source=source,
task_type=task_type,
arena=arena,
category=category,
max_results=max_results,
)
req = self._build_request(
method="GET",
path="/benchmarks",
base_url=base_url,
url_variables=url_variables,
request=request,
request_body_required=False,
request_has_path_params=False,
request_has_query_params=True,
user_agent_header="user-agent",
accept_header_value="application/json",
http_headers=http_headers,
_globals=operations.GetBenchmarksGlobals(
http_referer=self.sdk_configuration.globals.http_referer,
x_open_router_title=self.sdk_configuration.globals.x_open_router_title,
x_open_router_categories=self.sdk_configuration.globals.x_open_router_categories,
),
security=self.sdk_configuration.security,
allow_empty_value=None,
timeout_ms=timeout_ms,
)
if retries == UNSET:
if self.sdk_configuration.retry_config is not UNSET:
retries = self.sdk_configuration.retry_config
else:
retries = utils.RetryConfig(
"backoff", utils.BackoffStrategy(500, 60000, 1.5, 3600000), True
)
retry_config = None
if isinstance(retries, utils.RetryConfig):
retry_config = (retries, ["5XX"])
http_res = self.do_request(
hook_ctx=HookContext(
config=self.sdk_configuration,
base_url=base_url or "",
operation_id="getBenchmarks",
oauth2_scopes=None,
security_source=get_security_from_env(
self.sdk_configuration.security, components.Security
),
),
request=req,
error_status_codes=["400", "401", "429", "4XX", "500", "5XX"],
retry_config=retry_config,
)
response_data: Any = None
if utils.match_response(http_res, "200", "application/json"):
return unmarshal_json_response(
components.UnifiedBenchmarksResponse, http_res
)
if utils.match_response(http_res, "400", "application/json"):
response_data = unmarshal_json_response(
errors.BadRequestResponseErrorData, http_res
)
raise errors.BadRequestResponseError(response_data, http_res)
if utils.match_response(http_res, "401", "application/json"):
response_data = unmarshal_json_response(
errors.UnauthorizedResponseErrorData, http_res
)
raise errors.UnauthorizedResponseError(response_data, http_res)
if utils.match_response(http_res, "429", "application/json"):
response_data = unmarshal_json_response(
errors.TooManyRequestsResponseErrorData, http_res
)
raise errors.TooManyRequestsResponseError(response_data, http_res)
if utils.match_response(http_res, "500", "application/json"):
response_data = unmarshal_json_response(
errors.InternalServerResponseErrorData, http_res
)
raise errors.InternalServerResponseError(response_data, http_res)
if utils.match_response(http_res, "4XX", "*"):
http_res_text = utils.stream_to_text(http_res)
raise errors.OpenRouterDefaultError(
"API error occurred", http_res, http_res_text
)
if utils.match_response(http_res, "5XX", "*"):
http_res_text = utils.stream_to_text(http_res)
raise errors.OpenRouterDefaultError(
"API error occurred", http_res, http_res_text
)
raise errors.OpenRouterDefaultError("Unexpected response received", http_res)
async def get_benchmarks_async(
self,
*,
http_referer: Optional[str] = None,
x_open_router_title: Optional[str] = None,
x_open_router_categories: Optional[str] = None,
source: Optional[operations.Source] = None,
task_type: Optional[operations.TaskType] = None,
arena: Optional[operations.Arena] = None,
category: Optional[str] = None,
max_results: Optional[int] = None,
retries: OptionalNullable[utils.RetryConfig] = UNSET,
server_url: Optional[str] = None,
timeout_ms: Optional[int] = None,
http_headers: Optional[Mapping[str, str]] = None,
) -> components.UnifiedBenchmarksResponse:
r"""List Benchmarks
Unified benchmark endpoint that aggregates scores from multiple benchmark sources (Artificial Analysis, Design Arena). Filter by source to reproduce the exact shapes from the legacy per-source endpoints, or use task_type to find models suited for specific workloads. Authenticate with any valid OpenRouter API key. Rate-limited to 30 requests/minute per key and 500 requests/day per account.
:param http_referer: The app identifier should be your app's URL and is used as the primary identifier for rankings.
This is used to track API usage per application.
:param x_open_router_title: The app display name allows you to customize how your app appears in OpenRouter's dashboard.
:param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings.
:param source: Benchmark source to query. Determines the shape of the returned items. When omitted, returns results from all sources.
:param task_type: Filter results by task type. For Artificial Analysis, maps to the corresponding index. For Design Arena, maps to the matching category.
:param arena: Design Arena only: arena to query. Defaults to `models` when source is `design-arena`.
:param category: Design Arena only: category within the arena (e.g. `codecategories`, `uicomponent`, `gamedev`, `3d`, `dataviz`, `image`, `video`, `svg`). When omitted, returns all categories.
:param max_results: Maximum number of items to return. When omitted, all matching results are returned.
:param retries: Override the default retry configuration for this method
:param server_url: Override the default server URL for this method
:param timeout_ms: Override the default request timeout configuration for this method in milliseconds
:param http_headers: Additional headers to set or replace on requests.
"""
base_url = None
url_variables = None
if timeout_ms is None:
timeout_ms = self.sdk_configuration.timeout_ms
if server_url is not None:
base_url = server_url
else:
base_url = self._get_url(base_url, url_variables)
request = operations.GetBenchmarksRequest(
http_referer=http_referer,
x_open_router_title=x_open_router_title,
x_open_router_categories=x_open_router_categories,
source=source,
task_type=task_type,
arena=arena,
category=category,
max_results=max_results,
)
req = self._build_request_async(
method="GET",
path="/benchmarks",
base_url=base_url,
url_variables=url_variables,
request=request,
request_body_required=False,
request_has_path_params=False,
request_has_query_params=True,
user_agent_header="user-agent",
accept_header_value="application/json",
http_headers=http_headers,
_globals=operations.GetBenchmarksGlobals(
http_referer=self.sdk_configuration.globals.http_referer,
x_open_router_title=self.sdk_configuration.globals.x_open_router_title,
x_open_router_categories=self.sdk_configuration.globals.x_open_router_categories,
),
security=self.sdk_configuration.security,
allow_empty_value=None,
timeout_ms=timeout_ms,
)
if retries == UNSET:
if self.sdk_configuration.retry_config is not UNSET:
retries = self.sdk_configuration.retry_config
else:
retries = utils.RetryConfig(
"backoff", utils.BackoffStrategy(500, 60000, 1.5, 3600000), True
)
retry_config = None
if isinstance(retries, utils.RetryConfig):
retry_config = (retries, ["5XX"])
http_res = await self.do_request_async(
hook_ctx=HookContext(
config=self.sdk_configuration,
base_url=base_url or "",
operation_id="getBenchmarks",
oauth2_scopes=None,
security_source=get_security_from_env(
self.sdk_configuration.security, components.Security
),
),
request=req,
error_status_codes=["400", "401", "429", "4XX", "500", "5XX"],
retry_config=retry_config,
)
response_data: Any = None
if utils.match_response(http_res, "200", "application/json"):
return unmarshal_json_response(
components.UnifiedBenchmarksResponse, http_res
)
if utils.match_response(http_res, "400", "application/json"):
response_data = unmarshal_json_response(
errors.BadRequestResponseErrorData, http_res
)
raise errors.BadRequestResponseError(response_data, http_res)
if utils.match_response(http_res, "401", "application/json"):
response_data = unmarshal_json_response(
errors.UnauthorizedResponseErrorData, http_res
)
raise errors.UnauthorizedResponseError(response_data, http_res)
if utils.match_response(http_res, "429", "application/json"):
response_data = unmarshal_json_response(
errors.TooManyRequestsResponseErrorData, http_res
)
raise errors.TooManyRequestsResponseError(response_data, http_res)
if utils.match_response(http_res, "500", "application/json"):
response_data = unmarshal_json_response(
errors.InternalServerResponseErrorData, http_res
)
raise errors.InternalServerResponseError(response_data, http_res)
if utils.match_response(http_res, "4XX", "*"):
http_res_text = await utils.stream_to_text_async(http_res)
raise errors.OpenRouterDefaultError(
"API error occurred", http_res, http_res_text
)
if utils.match_response(http_res, "5XX", "*"):
http_res_text = await utils.stream_to_text_async(http_res)
raise errors.OpenRouterDefaultError(
"API error occurred", http_res, http_res_text
)
raise errors.OpenRouterDefaultError("Unexpected response received", http_res)
+2 -2
View File
@@ -302,7 +302,7 @@ class BetaAnalytics(BaseSDK):
:param dimensions:
:param filters:
:param granularity: Time granularity
:param group_limit: Maximum rows per distinct combination of dimensions (ClickHouse LIMIT n BY). When omitted on time-series queries (granularity + dimensions), auto-computed to avoid truncating time windows. Explicit values override the default and may truncate time buckets if set lower than the number of buckets in the range. Ignored when no dimensions are specified.
:param group_limit: Maximum rows per distinct combination of dimensions. When omitted on time-series queries (granularity + dimensions), auto-computed to avoid truncating time windows. Explicit values override the default and may truncate time buckets if set lower than the number of buckets in the range. Ignored when no dimensions are specified.
:param limit: Maximum total rows returned. Defaults to 1000. On time-series queries with dimensions and no explicit group_limit, the server may raise this to accommodate the expected number of unique time-bucket/dimension combinations.
:param order_by:
:param time_range:
@@ -480,7 +480,7 @@ class BetaAnalytics(BaseSDK):
:param dimensions:
:param filters:
:param granularity: Time granularity
:param group_limit: Maximum rows per distinct combination of dimensions (ClickHouse LIMIT n BY). When omitted on time-series queries (granularity + dimensions), auto-computed to avoid truncating time windows. Explicit values override the default and may truncate time buckets if set lower than the number of buckets in the range. Ignored when no dimensions are specified.
:param group_limit: Maximum rows per distinct combination of dimensions. When omitted on time-series queries (granularity + dimensions), auto-computed to avoid truncating time windows. Explicit values override the default and may truncate time buckets if set lower than the number of buckets in the range. Ignored when no dimensions are specified.
:param limit: Maximum total rows returned. Defaults to 1000. On time-series queries with dimensions and no explicit group_limit, the server may raise this to accommodate the expected number of unique time-bucket/dimension combinations.
:param order_by:
:param time_range:
+2 -2
View File
@@ -359,7 +359,7 @@ class Byok(BaseSDK):
) -> components.CreateBYOKKeyResponse:
r"""Create a BYOK provider credential
Create a new bring-your-own-key (BYOK) provider credential. The raw key is encrypted at rest and never returned in API responses. Defaults to the authenticated entity's default workspace; use the `workspace_id` body field to scope to a different workspace. [Management key](/docs/guides/overview/auth/management-api-keys) required.
Create a new bring-your-own-key (BYOK) provider credential. The raw key is encrypted at rest and never returned in API responses. Defaults to the authenticated entity's default workspace; use the `workspace_id` body field to scope to a different workspace. Treat the raw key as write-only; it is never returned after creation. [Management key](/docs/guides/overview/auth/management-api-keys) required.
:param key: The raw provider API key or credential. This value is encrypted at rest and never returned in API responses.
:param provider: The upstream provider this credential authenticates against, as a lowercase slug (e.g. `openai`, `anthropic`, `amazon-bedrock`).
@@ -520,7 +520,7 @@ class Byok(BaseSDK):
) -> components.CreateBYOKKeyResponse:
r"""Create a BYOK provider credential
Create a new bring-your-own-key (BYOK) provider credential. The raw key is encrypted at rest and never returned in API responses. Defaults to the authenticated entity's default workspace; use the `workspace_id` body field to scope to a different workspace. [Management key](/docs/guides/overview/auth/management-api-keys) required.
Create a new bring-your-own-key (BYOK) provider credential. The raw key is encrypted at rest and never returned in API responses. Defaults to the authenticated entity's default workspace; use the `workspace_id` body field to scope to a different workspace. Treat the raw key as write-only; it is never returned after creation. [Management key](/docs/guides/overview/auth/management-api-keys) required.
:param key: The raw provider API key or credential. This value is encrypted at rest and never returned in API responses.
:param provider: The upstream provider this credential authenticates against, as a lowercase slug (e.g. `openai`, `anthropic`, `amazon-bedrock`).
+82
View File
@@ -48,6 +48,7 @@ class Chat(BaseSDK):
max_completion_tokens: OptionalNullable[int] = UNSET,
max_tokens: OptionalNullable[int] = UNSET,
metadata: Optional[Dict[str, str]] = None,
min_p: OptionalNullable[float] = UNSET,
modalities: Optional[List[components.Modality]] = None,
model: Optional[str] = None,
models: Optional[List[str]] = None,
@@ -70,6 +71,10 @@ class Chat(BaseSDK):
components.ChatRequestReasoningTypedDict,
]
] = None,
reasoning_effort: OptionalNullable[
components.ChatRequestReasoningEffort
] = UNSET,
repetition_penalty: OptionalNullable[float] = UNSET,
response_format: Optional[
Union[components.ResponseFormat, components.ResponseFormatTypedDict]
] = None,
@@ -99,6 +104,8 @@ class Chat(BaseSDK):
List[components.ChatFunctionToolTypedDict],
]
] = None,
top_a: OptionalNullable[float] = UNSET,
top_k: OptionalNullable[int] = UNSET,
top_logprobs: OptionalNullable[int] = UNSET,
top_p: OptionalNullable[float] = UNSET,
trace: Optional[
@@ -132,6 +139,7 @@ class Chat(BaseSDK):
:param max_completion_tokens: Maximum tokens in completion
:param max_tokens: Maximum tokens (deprecated, use max_completion_tokens). Note: some providers enforce a minimum of 16.
:param metadata: Key-value pairs for additional object information (max 16 pairs, 64 char keys, 512 char values)
:param min_p: Minimum probability threshold relative to the most likely token. Tokens with probability below min_p * (probability of top token) are filtered out. Not all providers support this parameter.
:param modalities: Output modalities for the response. Supported values are \"text\", \"image\", and \"audio\".
:param model: Model to use for completion
:param models: Models to use for completion
@@ -140,6 +148,8 @@ class Chat(BaseSDK):
:param presence_penalty: Presence penalty (-2.0 to 2.0)
:param provider: When multiple model providers are available, optionally indicate your routing preference.
: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 repetition_penalty: Penalizes tokens based on how much they have already appeared in the text. A value of 1.0 means no penalty. Values above 1.0 penalize repeated tokens more strongly. Not all providers support this parameter.
:param response_format: Response format configuration
:param seed: Random seed for deterministic outputs
:param service_tier: The service tier to use for processing this request.
@@ -151,6 +161,8 @@ class Chat(BaseSDK):
:param temperature: Sampling temperature (0-2)
:param tool_choice: Tool choice configuration
:param tools: Available tools for function calling
:param top_a: Consider only tokens with \"sufficiently high\" probabilities based on the probability of the most likely token. Not all providers support this parameter.
:param top_k: Limits the model to choose from the top K most likely tokens at each step. A value of 1 means the model will always pick the most likely next token. Not all providers support this parameter.
:param top_logprobs: Number of top log probabilities to return (0-20)
:param top_p: Nucleus sampling parameter (0-1)
:param trace: Metadata for observability and tracing. Known keys (trace_id, trace_name, span_name, generation_name, parent_span_id) have special handling. Additional keys are passed through as custom metadata to configured broadcast destinations.
@@ -194,6 +206,7 @@ class Chat(BaseSDK):
max_completion_tokens: OptionalNullable[int] = UNSET,
max_tokens: OptionalNullable[int] = UNSET,
metadata: Optional[Dict[str, str]] = None,
min_p: OptionalNullable[float] = UNSET,
modalities: Optional[List[components.Modality]] = None,
model: Optional[str] = None,
models: Optional[List[str]] = None,
@@ -216,6 +229,10 @@ class Chat(BaseSDK):
components.ChatRequestReasoningTypedDict,
]
] = None,
reasoning_effort: OptionalNullable[
components.ChatRequestReasoningEffort
] = UNSET,
repetition_penalty: OptionalNullable[float] = UNSET,
response_format: Optional[
Union[components.ResponseFormat, components.ResponseFormatTypedDict]
] = None,
@@ -245,6 +262,8 @@ class Chat(BaseSDK):
List[components.ChatFunctionToolTypedDict],
]
] = None,
top_a: OptionalNullable[float] = UNSET,
top_k: OptionalNullable[int] = UNSET,
top_logprobs: OptionalNullable[int] = UNSET,
top_p: OptionalNullable[float] = UNSET,
trace: Optional[
@@ -278,6 +297,7 @@ class Chat(BaseSDK):
:param max_completion_tokens: Maximum tokens in completion
:param max_tokens: Maximum tokens (deprecated, use max_completion_tokens). Note: some providers enforce a minimum of 16.
:param metadata: Key-value pairs for additional object information (max 16 pairs, 64 char keys, 512 char values)
:param min_p: Minimum probability threshold relative to the most likely token. Tokens with probability below min_p * (probability of top token) are filtered out. Not all providers support this parameter.
:param modalities: Output modalities for the response. Supported values are \"text\", \"image\", and \"audio\".
:param model: Model to use for completion
:param models: Models to use for completion
@@ -286,6 +306,8 @@ class Chat(BaseSDK):
:param presence_penalty: Presence penalty (-2.0 to 2.0)
:param provider: When multiple model providers are available, optionally indicate your routing preference.
: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 repetition_penalty: Penalizes tokens based on how much they have already appeared in the text. A value of 1.0 means no penalty. Values above 1.0 penalize repeated tokens more strongly. Not all providers support this parameter.
:param response_format: Response format configuration
:param seed: Random seed for deterministic outputs
:param service_tier: The service tier to use for processing this request.
@@ -297,6 +319,8 @@ class Chat(BaseSDK):
:param temperature: Sampling temperature (0-2)
:param tool_choice: Tool choice configuration
:param tools: Available tools for function calling
:param top_a: Consider only tokens with \"sufficiently high\" probabilities based on the probability of the most likely token. Not all providers support this parameter.
:param top_k: Limits the model to choose from the top K most likely tokens at each step. A value of 1 means the model will always pick the most likely next token. Not all providers support this parameter.
:param top_logprobs: Number of top log probabilities to return (0-20)
:param top_p: Nucleus sampling parameter (0-1)
:param trace: Metadata for observability and tracing. Known keys (trace_id, trace_name, span_name, generation_name, parent_span_id) have special handling. Additional keys are passed through as custom metadata to configured broadcast destinations.
@@ -339,6 +363,7 @@ class Chat(BaseSDK):
max_completion_tokens: OptionalNullable[int] = UNSET,
max_tokens: OptionalNullable[int] = UNSET,
metadata: Optional[Dict[str, str]] = None,
min_p: OptionalNullable[float] = UNSET,
modalities: Optional[List[components.Modality]] = None,
model: Optional[str] = None,
models: Optional[List[str]] = None,
@@ -361,6 +386,10 @@ class Chat(BaseSDK):
components.ChatRequestReasoningTypedDict,
]
] = None,
reasoning_effort: OptionalNullable[
components.ChatRequestReasoningEffort
] = UNSET,
repetition_penalty: OptionalNullable[float] = UNSET,
response_format: Optional[
Union[components.ResponseFormat, components.ResponseFormatTypedDict]
] = None,
@@ -390,6 +419,8 @@ class Chat(BaseSDK):
List[components.ChatFunctionToolTypedDict],
]
] = None,
top_a: OptionalNullable[float] = UNSET,
top_k: OptionalNullable[int] = UNSET,
top_logprobs: OptionalNullable[int] = UNSET,
top_p: OptionalNullable[float] = UNSET,
trace: Optional[
@@ -423,6 +454,7 @@ class Chat(BaseSDK):
:param max_completion_tokens: Maximum tokens in completion
:param max_tokens: Maximum tokens (deprecated, use max_completion_tokens). Note: some providers enforce a minimum of 16.
:param metadata: Key-value pairs for additional object information (max 16 pairs, 64 char keys, 512 char values)
:param min_p: Minimum probability threshold relative to the most likely token. Tokens with probability below min_p * (probability of top token) are filtered out. Not all providers support this parameter.
:param modalities: Output modalities for the response. Supported values are \"text\", \"image\", and \"audio\".
:param model: Model to use for completion
:param models: Models to use for completion
@@ -431,6 +463,8 @@ class Chat(BaseSDK):
:param presence_penalty: Presence penalty (-2.0 to 2.0)
:param provider: When multiple model providers are available, optionally indicate your routing preference.
: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 repetition_penalty: Penalizes tokens based on how much they have already appeared in the text. A value of 1.0 means no penalty. Values above 1.0 penalize repeated tokens more strongly. Not all providers support this parameter.
:param response_format: Response format configuration
:param seed: Random seed for deterministic outputs
:param service_tier: The service tier to use for processing this request.
@@ -442,6 +476,8 @@ class Chat(BaseSDK):
:param temperature: Sampling temperature (0-2)
:param tool_choice: Tool choice configuration
:param tools: Available tools for function calling
:param top_a: Consider only tokens with \"sufficiently high\" probabilities based on the probability of the most likely token. Not all providers support this parameter.
:param top_k: Limits the model to choose from the top K most likely tokens at each step. A value of 1 means the model will always pick the most likely next token. Not all providers support this parameter.
:param top_logprobs: Number of top log probabilities to return (0-20)
:param top_p: Nucleus sampling parameter (0-1)
:param trace: Metadata for observability and tracing. Known keys (trace_id, trace_name, span_name, generation_name, parent_span_id) have special handling. Additional keys are passed through as custom metadata to configured broadcast destinations.
@@ -484,6 +520,7 @@ class Chat(BaseSDK):
messages, List[components.ChatMessages]
),
metadata=metadata,
min_p=min_p,
modalities=modalities,
model=model,
models=models,
@@ -498,6 +535,8 @@ class Chat(BaseSDK):
reasoning=utils.get_pydantic_model(
reasoning, Optional[components.ChatRequestReasoning]
),
reasoning_effort=reasoning_effort,
repetition_penalty=repetition_penalty,
response_format=utils.get_pydantic_model(
response_format, Optional[components.ResponseFormat]
),
@@ -520,6 +559,8 @@ class Chat(BaseSDK):
tools=utils.get_pydantic_model(
tools, Optional[List[components.ChatFunctionTool]]
),
top_a=top_a,
top_k=top_k,
top_logprobs=top_logprobs,
top_p=top_p,
trace=utils.get_pydantic_model(trace, Optional[components.TraceConfig]),
@@ -764,6 +805,7 @@ class Chat(BaseSDK):
max_completion_tokens: OptionalNullable[int] = UNSET,
max_tokens: OptionalNullable[int] = UNSET,
metadata: Optional[Dict[str, str]] = None,
min_p: OptionalNullable[float] = UNSET,
modalities: Optional[List[components.Modality]] = None,
model: Optional[str] = None,
models: Optional[List[str]] = None,
@@ -786,6 +828,10 @@ class Chat(BaseSDK):
components.ChatRequestReasoningTypedDict,
]
] = None,
reasoning_effort: OptionalNullable[
components.ChatRequestReasoningEffort
] = UNSET,
repetition_penalty: OptionalNullable[float] = UNSET,
response_format: Optional[
Union[components.ResponseFormat, components.ResponseFormatTypedDict]
] = None,
@@ -815,6 +861,8 @@ class Chat(BaseSDK):
List[components.ChatFunctionToolTypedDict],
]
] = None,
top_a: OptionalNullable[float] = UNSET,
top_k: OptionalNullable[int] = UNSET,
top_logprobs: OptionalNullable[int] = UNSET,
top_p: OptionalNullable[float] = UNSET,
trace: Optional[
@@ -848,6 +896,7 @@ class Chat(BaseSDK):
:param max_completion_tokens: Maximum tokens in completion
:param max_tokens: Maximum tokens (deprecated, use max_completion_tokens). Note: some providers enforce a minimum of 16.
:param metadata: Key-value pairs for additional object information (max 16 pairs, 64 char keys, 512 char values)
:param min_p: Minimum probability threshold relative to the most likely token. Tokens with probability below min_p * (probability of top token) are filtered out. Not all providers support this parameter.
:param modalities: Output modalities for the response. Supported values are \"text\", \"image\", and \"audio\".
:param model: Model to use for completion
:param models: Models to use for completion
@@ -856,6 +905,8 @@ class Chat(BaseSDK):
:param presence_penalty: Presence penalty (-2.0 to 2.0)
:param provider: When multiple model providers are available, optionally indicate your routing preference.
: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 repetition_penalty: Penalizes tokens based on how much they have already appeared in the text. A value of 1.0 means no penalty. Values above 1.0 penalize repeated tokens more strongly. Not all providers support this parameter.
:param response_format: Response format configuration
:param seed: Random seed for deterministic outputs
:param service_tier: The service tier to use for processing this request.
@@ -867,6 +918,8 @@ class Chat(BaseSDK):
:param temperature: Sampling temperature (0-2)
:param tool_choice: Tool choice configuration
:param tools: Available tools for function calling
:param top_a: Consider only tokens with \"sufficiently high\" probabilities based on the probability of the most likely token. Not all providers support this parameter.
:param top_k: Limits the model to choose from the top K most likely tokens at each step. A value of 1 means the model will always pick the most likely next token. Not all providers support this parameter.
:param top_logprobs: Number of top log probabilities to return (0-20)
:param top_p: Nucleus sampling parameter (0-1)
:param trace: Metadata for observability and tracing. Known keys (trace_id, trace_name, span_name, generation_name, parent_span_id) have special handling. Additional keys are passed through as custom metadata to configured broadcast destinations.
@@ -910,6 +963,7 @@ class Chat(BaseSDK):
max_completion_tokens: OptionalNullable[int] = UNSET,
max_tokens: OptionalNullable[int] = UNSET,
metadata: Optional[Dict[str, str]] = None,
min_p: OptionalNullable[float] = UNSET,
modalities: Optional[List[components.Modality]] = None,
model: Optional[str] = None,
models: Optional[List[str]] = None,
@@ -932,6 +986,10 @@ class Chat(BaseSDK):
components.ChatRequestReasoningTypedDict,
]
] = None,
reasoning_effort: OptionalNullable[
components.ChatRequestReasoningEffort
] = UNSET,
repetition_penalty: OptionalNullable[float] = UNSET,
response_format: Optional[
Union[components.ResponseFormat, components.ResponseFormatTypedDict]
] = None,
@@ -961,6 +1019,8 @@ class Chat(BaseSDK):
List[components.ChatFunctionToolTypedDict],
]
] = None,
top_a: OptionalNullable[float] = UNSET,
top_k: OptionalNullable[int] = UNSET,
top_logprobs: OptionalNullable[int] = UNSET,
top_p: OptionalNullable[float] = UNSET,
trace: Optional[
@@ -994,6 +1054,7 @@ class Chat(BaseSDK):
:param max_completion_tokens: Maximum tokens in completion
:param max_tokens: Maximum tokens (deprecated, use max_completion_tokens). Note: some providers enforce a minimum of 16.
:param metadata: Key-value pairs for additional object information (max 16 pairs, 64 char keys, 512 char values)
:param min_p: Minimum probability threshold relative to the most likely token. Tokens with probability below min_p * (probability of top token) are filtered out. Not all providers support this parameter.
:param modalities: Output modalities for the response. Supported values are \"text\", \"image\", and \"audio\".
:param model: Model to use for completion
:param models: Models to use for completion
@@ -1002,6 +1063,8 @@ class Chat(BaseSDK):
:param presence_penalty: Presence penalty (-2.0 to 2.0)
:param provider: When multiple model providers are available, optionally indicate your routing preference.
: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 repetition_penalty: Penalizes tokens based on how much they have already appeared in the text. A value of 1.0 means no penalty. Values above 1.0 penalize repeated tokens more strongly. Not all providers support this parameter.
:param response_format: Response format configuration
:param seed: Random seed for deterministic outputs
:param service_tier: The service tier to use for processing this request.
@@ -1013,6 +1076,8 @@ class Chat(BaseSDK):
:param temperature: Sampling temperature (0-2)
:param tool_choice: Tool choice configuration
:param tools: Available tools for function calling
:param top_a: Consider only tokens with \"sufficiently high\" probabilities based on the probability of the most likely token. Not all providers support this parameter.
:param top_k: Limits the model to choose from the top K most likely tokens at each step. A value of 1 means the model will always pick the most likely next token. Not all providers support this parameter.
:param top_logprobs: Number of top log probabilities to return (0-20)
:param top_p: Nucleus sampling parameter (0-1)
:param trace: Metadata for observability and tracing. Known keys (trace_id, trace_name, span_name, generation_name, parent_span_id) have special handling. Additional keys are passed through as custom metadata to configured broadcast destinations.
@@ -1055,6 +1120,7 @@ class Chat(BaseSDK):
max_completion_tokens: OptionalNullable[int] = UNSET,
max_tokens: OptionalNullable[int] = UNSET,
metadata: Optional[Dict[str, str]] = None,
min_p: OptionalNullable[float] = UNSET,
modalities: Optional[List[components.Modality]] = None,
model: Optional[str] = None,
models: Optional[List[str]] = None,
@@ -1077,6 +1143,10 @@ class Chat(BaseSDK):
components.ChatRequestReasoningTypedDict,
]
] = None,
reasoning_effort: OptionalNullable[
components.ChatRequestReasoningEffort
] = UNSET,
repetition_penalty: OptionalNullable[float] = UNSET,
response_format: Optional[
Union[components.ResponseFormat, components.ResponseFormatTypedDict]
] = None,
@@ -1106,6 +1176,8 @@ class Chat(BaseSDK):
List[components.ChatFunctionToolTypedDict],
]
] = None,
top_a: OptionalNullable[float] = UNSET,
top_k: OptionalNullable[int] = UNSET,
top_logprobs: OptionalNullable[int] = UNSET,
top_p: OptionalNullable[float] = UNSET,
trace: Optional[
@@ -1139,6 +1211,7 @@ class Chat(BaseSDK):
:param max_completion_tokens: Maximum tokens in completion
:param max_tokens: Maximum tokens (deprecated, use max_completion_tokens). Note: some providers enforce a minimum of 16.
:param metadata: Key-value pairs for additional object information (max 16 pairs, 64 char keys, 512 char values)
:param min_p: Minimum probability threshold relative to the most likely token. Tokens with probability below min_p * (probability of top token) are filtered out. Not all providers support this parameter.
:param modalities: Output modalities for the response. Supported values are \"text\", \"image\", and \"audio\".
:param model: Model to use for completion
:param models: Models to use for completion
@@ -1147,6 +1220,8 @@ class Chat(BaseSDK):
:param presence_penalty: Presence penalty (-2.0 to 2.0)
:param provider: When multiple model providers are available, optionally indicate your routing preference.
: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 repetition_penalty: Penalizes tokens based on how much they have already appeared in the text. A value of 1.0 means no penalty. Values above 1.0 penalize repeated tokens more strongly. Not all providers support this parameter.
:param response_format: Response format configuration
:param seed: Random seed for deterministic outputs
:param service_tier: The service tier to use for processing this request.
@@ -1158,6 +1233,8 @@ class Chat(BaseSDK):
:param temperature: Sampling temperature (0-2)
:param tool_choice: Tool choice configuration
:param tools: Available tools for function calling
:param top_a: Consider only tokens with \"sufficiently high\" probabilities based on the probability of the most likely token. Not all providers support this parameter.
:param top_k: Limits the model to choose from the top K most likely tokens at each step. A value of 1 means the model will always pick the most likely next token. Not all providers support this parameter.
:param top_logprobs: Number of top log probabilities to return (0-20)
:param top_p: Nucleus sampling parameter (0-1)
:param trace: Metadata for observability and tracing. Known keys (trace_id, trace_name, span_name, generation_name, parent_span_id) have special handling. Additional keys are passed through as custom metadata to configured broadcast destinations.
@@ -1200,6 +1277,7 @@ class Chat(BaseSDK):
messages, List[components.ChatMessages]
),
metadata=metadata,
min_p=min_p,
modalities=modalities,
model=model,
models=models,
@@ -1214,6 +1292,8 @@ class Chat(BaseSDK):
reasoning=utils.get_pydantic_model(
reasoning, Optional[components.ChatRequestReasoning]
),
reasoning_effort=reasoning_effort,
repetition_penalty=repetition_penalty,
response_format=utils.get_pydantic_model(
response_format, Optional[components.ResponseFormat]
),
@@ -1236,6 +1316,8 @@ class Chat(BaseSDK):
tools=utils.get_pydantic_model(
tools, Optional[List[components.ChatFunctionTool]]
),
top_a=top_a,
top_k=top_k,
top_logprobs=top_logprobs,
top_p=top_p,
trace=utils.get_pydantic_model(trace, Optional[components.TraceConfig]),
+315
View File
@@ -0,0 +1,315 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from .basesdk import BaseSDK
from openrouter import components, errors, operations, utils
from openrouter._hooks import HookContext
from openrouter.types import OptionalNullable, UNSET
from openrouter.utils import get_security_from_env
from openrouter.utils.unmarshal_json_response import unmarshal_json_response
from typing import Any, Mapping, Optional
class Classifications(BaseSDK):
r"""Task classification market-share endpoints"""
def get_task_classifications(
self,
*,
http_referer: Optional[str] = None,
x_open_router_title: Optional[str] = None,
x_open_router_categories: Optional[str] = None,
window: Optional[operations.Window] = "7d",
retries: OptionalNullable[utils.RetryConfig] = UNSET,
server_url: Optional[str] = None,
timeout_ms: Optional[int] = None,
http_headers: Optional[Mapping[str, str]] = None,
) -> components.TaskClassificationResponse:
r"""Task classification market share
Returns the market-share breakdown of OpenRouter traffic by task classification
(e.g. code generation, web search, summarization) over a trailing time window.
Each classification reports its share of classified sampled requests (`usage_share`)
and classified sampled token volume (`token_share`) as fractions between 0 and 1.
The unclassified `other` bucket is excluded. Absolute volumes are not exposed
because the underlying data is sampled.
Each classification also includes a `models` array listing the top models by
request volume within that classification, with their within-tag usage and token shares.
Classifications are grouped into macro-categories (Code, Data, Agent, General)
with aggregate shares provided for each.
Authenticate with any valid OpenRouter API key (same key used for inference).
Rate-limited to 30 requests/minute per key and 500 requests/day per account.
When republishing or quoting this data, cite as:
\"Source: OpenRouter (openrouter.ai/rankings), as of {as_of}.\"
:param http_referer: The app identifier should be your app's URL and is used as the primary identifier for rankings.
This is used to track API usage per application.
:param x_open_router_title: The app display name allows you to customize how your app appears in OpenRouter's dashboard.
:param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings.
:param window: Trailing time window for the classification data. Currently only `7d` (trailing 7 days) is supported.
:param retries: Override the default retry configuration for this method
:param server_url: Override the default server URL for this method
:param timeout_ms: Override the default request timeout configuration for this method in milliseconds
:param http_headers: Additional headers to set or replace on requests.
"""
base_url = None
url_variables = None
if timeout_ms is None:
timeout_ms = self.sdk_configuration.timeout_ms
if server_url is not None:
base_url = server_url
else:
base_url = self._get_url(base_url, url_variables)
request = operations.GetTaskClassificationsRequest(
http_referer=http_referer,
x_open_router_title=x_open_router_title,
x_open_router_categories=x_open_router_categories,
window=window,
)
req = self._build_request(
method="GET",
path="/classifications/task",
base_url=base_url,
url_variables=url_variables,
request=request,
request_body_required=False,
request_has_path_params=False,
request_has_query_params=True,
user_agent_header="user-agent",
accept_header_value="application/json",
http_headers=http_headers,
_globals=operations.GetTaskClassificationsGlobals(
http_referer=self.sdk_configuration.globals.http_referer,
x_open_router_title=self.sdk_configuration.globals.x_open_router_title,
x_open_router_categories=self.sdk_configuration.globals.x_open_router_categories,
),
security=self.sdk_configuration.security,
allow_empty_value=None,
timeout_ms=timeout_ms,
)
if retries == UNSET:
if self.sdk_configuration.retry_config is not UNSET:
retries = self.sdk_configuration.retry_config
else:
retries = utils.RetryConfig(
"backoff", utils.BackoffStrategy(500, 60000, 1.5, 3600000), True
)
retry_config = None
if isinstance(retries, utils.RetryConfig):
retry_config = (retries, ["5XX"])
http_res = self.do_request(
hook_ctx=HookContext(
config=self.sdk_configuration,
base_url=base_url or "",
operation_id="getTaskClassifications",
oauth2_scopes=None,
security_source=get_security_from_env(
self.sdk_configuration.security, components.Security
),
),
request=req,
error_status_codes=["400", "401", "429", "4XX", "500", "5XX"],
retry_config=retry_config,
)
response_data: Any = None
if utils.match_response(http_res, "200", "application/json"):
return unmarshal_json_response(
components.TaskClassificationResponse, http_res
)
if utils.match_response(http_res, "400", "application/json"):
response_data = unmarshal_json_response(
errors.BadRequestResponseErrorData, http_res
)
raise errors.BadRequestResponseError(response_data, http_res)
if utils.match_response(http_res, "401", "application/json"):
response_data = unmarshal_json_response(
errors.UnauthorizedResponseErrorData, http_res
)
raise errors.UnauthorizedResponseError(response_data, http_res)
if utils.match_response(http_res, "429", "application/json"):
response_data = unmarshal_json_response(
errors.TooManyRequestsResponseErrorData, http_res
)
raise errors.TooManyRequestsResponseError(response_data, http_res)
if utils.match_response(http_res, "500", "application/json"):
response_data = unmarshal_json_response(
errors.InternalServerResponseErrorData, http_res
)
raise errors.InternalServerResponseError(response_data, http_res)
if utils.match_response(http_res, "4XX", "*"):
http_res_text = utils.stream_to_text(http_res)
raise errors.OpenRouterDefaultError(
"API error occurred", http_res, http_res_text
)
if utils.match_response(http_res, "5XX", "*"):
http_res_text = utils.stream_to_text(http_res)
raise errors.OpenRouterDefaultError(
"API error occurred", http_res, http_res_text
)
raise errors.OpenRouterDefaultError("Unexpected response received", http_res)
async def get_task_classifications_async(
self,
*,
http_referer: Optional[str] = None,
x_open_router_title: Optional[str] = None,
x_open_router_categories: Optional[str] = None,
window: Optional[operations.Window] = "7d",
retries: OptionalNullable[utils.RetryConfig] = UNSET,
server_url: Optional[str] = None,
timeout_ms: Optional[int] = None,
http_headers: Optional[Mapping[str, str]] = None,
) -> components.TaskClassificationResponse:
r"""Task classification market share
Returns the market-share breakdown of OpenRouter traffic by task classification
(e.g. code generation, web search, summarization) over a trailing time window.
Each classification reports its share of classified sampled requests (`usage_share`)
and classified sampled token volume (`token_share`) as fractions between 0 and 1.
The unclassified `other` bucket is excluded. Absolute volumes are not exposed
because the underlying data is sampled.
Each classification also includes a `models` array listing the top models by
request volume within that classification, with their within-tag usage and token shares.
Classifications are grouped into macro-categories (Code, Data, Agent, General)
with aggregate shares provided for each.
Authenticate with any valid OpenRouter API key (same key used for inference).
Rate-limited to 30 requests/minute per key and 500 requests/day per account.
When republishing or quoting this data, cite as:
\"Source: OpenRouter (openrouter.ai/rankings), as of {as_of}.\"
:param http_referer: The app identifier should be your app's URL and is used as the primary identifier for rankings.
This is used to track API usage per application.
:param x_open_router_title: The app display name allows you to customize how your app appears in OpenRouter's dashboard.
:param x_open_router_categories: Comma-separated list of app categories (e.g. \"cli-agent,cloud-agent\"). Used for marketplace rankings.
:param window: Trailing time window for the classification data. Currently only `7d` (trailing 7 days) is supported.
:param retries: Override the default retry configuration for this method
:param server_url: Override the default server URL for this method
:param timeout_ms: Override the default request timeout configuration for this method in milliseconds
:param http_headers: Additional headers to set or replace on requests.
"""
base_url = None
url_variables = None
if timeout_ms is None:
timeout_ms = self.sdk_configuration.timeout_ms
if server_url is not None:
base_url = server_url
else:
base_url = self._get_url(base_url, url_variables)
request = operations.GetTaskClassificationsRequest(
http_referer=http_referer,
x_open_router_title=x_open_router_title,
x_open_router_categories=x_open_router_categories,
window=window,
)
req = self._build_request_async(
method="GET",
path="/classifications/task",
base_url=base_url,
url_variables=url_variables,
request=request,
request_body_required=False,
request_has_path_params=False,
request_has_query_params=True,
user_agent_header="user-agent",
accept_header_value="application/json",
http_headers=http_headers,
_globals=operations.GetTaskClassificationsGlobals(
http_referer=self.sdk_configuration.globals.http_referer,
x_open_router_title=self.sdk_configuration.globals.x_open_router_title,
x_open_router_categories=self.sdk_configuration.globals.x_open_router_categories,
),
security=self.sdk_configuration.security,
allow_empty_value=None,
timeout_ms=timeout_ms,
)
if retries == UNSET:
if self.sdk_configuration.retry_config is not UNSET:
retries = self.sdk_configuration.retry_config
else:
retries = utils.RetryConfig(
"backoff", utils.BackoffStrategy(500, 60000, 1.5, 3600000), True
)
retry_config = None
if isinstance(retries, utils.RetryConfig):
retry_config = (retries, ["5XX"])
http_res = await self.do_request_async(
hook_ctx=HookContext(
config=self.sdk_configuration,
base_url=base_url or "",
operation_id="getTaskClassifications",
oauth2_scopes=None,
security_source=get_security_from_env(
self.sdk_configuration.security, components.Security
),
),
request=req,
error_status_codes=["400", "401", "429", "4XX", "500", "5XX"],
retry_config=retry_config,
)
response_data: Any = None
if utils.match_response(http_res, "200", "application/json"):
return unmarshal_json_response(
components.TaskClassificationResponse, http_res
)
if utils.match_response(http_res, "400", "application/json"):
response_data = unmarshal_json_response(
errors.BadRequestResponseErrorData, http_res
)
raise errors.BadRequestResponseError(response_data, http_res)
if utils.match_response(http_res, "401", "application/json"):
response_data = unmarshal_json_response(
errors.UnauthorizedResponseErrorData, http_res
)
raise errors.UnauthorizedResponseError(response_data, http_res)
if utils.match_response(http_res, "429", "application/json"):
response_data = unmarshal_json_response(
errors.TooManyRequestsResponseErrorData, http_res
)
raise errors.TooManyRequestsResponseError(response_data, http_res)
if utils.match_response(http_res, "500", "application/json"):
response_data = unmarshal_json_response(
errors.InternalServerResponseErrorData, http_res
)
raise errors.InternalServerResponseError(response_data, http_res)
if utils.match_response(http_res, "4XX", "*"):
http_res_text = await utils.stream_to_text_async(http_res)
raise errors.OpenRouterDefaultError(
"API error occurred", http_res, http_res_text
)
if utils.match_response(http_res, "5XX", "*"):
http_res_text = await utils.stream_to_text_async(http_res)
raise errors.OpenRouterDefaultError(
"API error occurred", http_res, http_res_text
)
raise errors.OpenRouterDefaultError("Unexpected response received", http_res)
File diff suppressed because it is too large Load Diff
@@ -0,0 +1,60 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import BaseModel, Nullable, UNSET_SENTINEL
from pydantic import model_serializer
from typing_extensions import TypedDict
class AABenchmarkEntryTypedDict(TypedDict):
r"""Artificial Analysis benchmark index scores."""
agentic_index: Nullable[float]
r"""Artificial Analysis Agentic Index score"""
coding_index: Nullable[float]
r"""Artificial Analysis Coding Index score"""
intelligence_index: Nullable[float]
r"""Artificial Analysis Intelligence Index score"""
class AABenchmarkEntry(BaseModel):
r"""Artificial Analysis benchmark index scores."""
agentic_index: Nullable[float]
r"""Artificial Analysis Agentic Index score"""
coding_index: Nullable[float]
r"""Artificial Analysis Coding Index score"""
intelligence_index: Nullable[float]
r"""Artificial Analysis Intelligence Index score"""
@model_serializer(mode="wrap")
def serialize_model(self, handler):
optional_fields = []
nullable_fields = ["agentic_index", "coding_index", "intelligence_index"]
null_default_fields = []
serialized = handler(self)
m = {}
for n, f in type(self).model_fields.items():
k = f.alias or n
val = serialized.get(k)
serialized.pop(k, None)
optional_nullable = k in optional_fields and k in nullable_fields
is_set = (
self.__pydantic_fields_set__.intersection({n})
or k in null_default_fields
) # pylint: disable=no-member
if val is not None and val != UNSET_SENTINEL:
m[k] = val
elif val != UNSET_SENTINEL and (
not k in optional_fields or (optional_nullable and is_set)
):
m[k] = val
return m
@@ -9,15 +9,14 @@ from typing_extensions import NotRequired, TypedDict
class AdvisorNestedToolTypedDict(TypedDict):
r"""A tool made available to the advisor sub-agent. Accepts function tools and OpenRouter server tools (e.g. openrouter:web_search). The advisor tool may not list itself."""
r"""A tool made available to the advisor sub-agent. Only OpenRouter server tools (e.g. openrouter:web_search) are supported; function tools are rejected because the advisor has no way to execute them. The advisor tool may not list itself."""
type: str
function: NotRequired[Dict[str, Nullable[Any]]]
parameters: NotRequired[Dict[str, Nullable[Any]]]
class AdvisorNestedTool(BaseModel):
r"""A tool made available to the advisor sub-agent. Accepts function tools and OpenRouter server tools (e.g. openrouter:web_search). The advisor tool may not list itself."""
r"""A tool made available to the advisor sub-agent. Only OpenRouter server tools (e.g. openrouter:web_search) are supported; function tools are rejected because the advisor has no way to execute them. The advisor tool may not list itself."""
model_config = ConfigDict(
populate_by_name=True, arbitrary_types_allowed=True, extra="allow"
@@ -26,8 +25,6 @@ class AdvisorNestedTool(BaseModel):
type: str
function: Optional[Dict[str, Nullable[Any]]] = None
parameters: Optional[Dict[str, Nullable[Any]]] = None
@property
@@ -10,6 +10,7 @@ from typing_extensions import Annotated, NotRequired, TypedDict
AdvisorReasoningEffort = Union[
Literal[
"max",
"xhigh",
"high",
"medium",
@@ -30,7 +30,7 @@ class AdvisorServerToolConfigTypedDict(TypedDict):
temperature: NotRequired[float]
r"""Sampling temperature forwarded to the advisor call. When omitted, the provider's default applies."""
tools: NotRequired[List[AdvisorNestedToolTypedDict]]
r"""Tools the advisor sub-agent may use while forming its advice. The advisor runs as an agentic sub-agent over these tools, then returns its text. Must not include the advisor tool itself."""
r"""Tools the advisor sub-agent may use while forming its advice. The advisor runs as an agentic sub-agent over these tools, then returns its text. Only OpenRouter server tools are supported — function tools are rejected — and the list must not include the advisor tool itself."""
class AdvisorServerToolConfig(BaseModel):
@@ -64,4 +64,4 @@ class AdvisorServerToolConfig(BaseModel):
r"""Sampling temperature forwarded to the advisor call. When omitted, the provider's default applies."""
tools: Optional[List[AdvisorNestedTool]] = None
r"""Tools the advisor sub-agent may use while forming its advice. The advisor runs as an agentic sub-agent over these tools, then returns its text. Must not include the advisor tool itself."""
r"""Tools the advisor sub-agent may use while forming its advice. The advisor runs as an agentic sub-agent over these tools, then returns its text. Only OpenRouter server tools are supported — function tools are rejected — and the list must not include the advisor tool itself."""
@@ -0,0 +1,82 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .anthropiciterationcachecreation import (
AnthropicIterationCacheCreation,
AnthropicIterationCacheCreationTypedDict,
)
from openrouter.types import (
BaseModel,
Nullable,
OptionalNullable,
UNSET,
UNSET_SENTINEL,
)
from pydantic import model_serializer
from typing import Literal, Optional
from typing_extensions import NotRequired, TypedDict
AnthropicAdvisorMessageUsageIterationType = Literal["advisor_message",]
class AnthropicAdvisorMessageUsageIterationTypedDict(TypedDict):
model: str
type: AnthropicAdvisorMessageUsageIterationType
cache_creation: NotRequired[Nullable[AnthropicIterationCacheCreationTypedDict]]
cache_creation_input_tokens: NotRequired[int]
cache_read_input_tokens: NotRequired[int]
input_tokens: NotRequired[int]
output_tokens: NotRequired[int]
class AnthropicAdvisorMessageUsageIteration(BaseModel):
model: str
type: AnthropicAdvisorMessageUsageIterationType
cache_creation: OptionalNullable[AnthropicIterationCacheCreation] = UNSET
cache_creation_input_tokens: Optional[int] = None
cache_read_input_tokens: Optional[int] = None
input_tokens: Optional[int] = None
output_tokens: Optional[int] = None
@model_serializer(mode="wrap")
def serialize_model(self, handler):
optional_fields = [
"cache_creation",
"cache_creation_input_tokens",
"cache_read_input_tokens",
"input_tokens",
"output_tokens",
]
nullable_fields = ["cache_creation"]
null_default_fields = []
serialized = handler(self)
m = {}
for n, f in type(self).model_fields.items():
k = f.alias or n
val = serialized.get(k)
serialized.pop(k, None)
optional_nullable = k in optional_fields and k in nullable_fields
is_set = (
self.__pydantic_fields_set__.intersection({n})
or k in null_default_fields
) # pylint: disable=no-member
if val is not None and val != UNSET_SENTINEL:
m[k] = val
elif val != UNSET_SENTINEL and (
not k in optional_fields or (optional_nullable and is_set)
):
m[k] = val
return m
@@ -0,0 +1,79 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .anthropiciterationcachecreation import (
AnthropicIterationCacheCreation,
AnthropicIterationCacheCreationTypedDict,
)
from openrouter.types import (
BaseModel,
Nullable,
OptionalNullable,
UNSET,
UNSET_SENTINEL,
)
from pydantic import model_serializer
from typing import Literal, Optional
from typing_extensions import NotRequired, TypedDict
AnthropicCompactionUsageIterationType = Literal["compaction",]
class AnthropicCompactionUsageIterationTypedDict(TypedDict):
type: AnthropicCompactionUsageIterationType
cache_creation: NotRequired[Nullable[AnthropicIterationCacheCreationTypedDict]]
cache_creation_input_tokens: NotRequired[int]
cache_read_input_tokens: NotRequired[int]
input_tokens: NotRequired[int]
output_tokens: NotRequired[int]
class AnthropicCompactionUsageIteration(BaseModel):
type: AnthropicCompactionUsageIterationType
cache_creation: OptionalNullable[AnthropicIterationCacheCreation] = UNSET
cache_creation_input_tokens: Optional[int] = None
cache_read_input_tokens: Optional[int] = None
input_tokens: Optional[int] = None
output_tokens: Optional[int] = None
@model_serializer(mode="wrap")
def serialize_model(self, handler):
optional_fields = [
"cache_creation",
"cache_creation_input_tokens",
"cache_read_input_tokens",
"input_tokens",
"output_tokens",
]
nullable_fields = ["cache_creation"]
null_default_fields = []
serialized = handler(self)
m = {}
for n, f in type(self).model_fields.items():
k = f.alias or n
val = serialized.get(k)
serialized.pop(k, None)
optional_nullable = k in optional_fields and k in nullable_fields
is_set = (
self.__pydantic_fields_set__.intersection({n})
or k in null_default_fields
) # pylint: disable=no-member
if val is not None and val != UNSET_SENTINEL:
m[k] = val
elif val != UNSET_SENTINEL and (
not k in optional_fields or (optional_nullable and is_set)
):
m[k] = val
return m
@@ -9,6 +9,10 @@ from .anthropiccachecontroldirective import (
AnthropicCacheControlDirective,
AnthropicCacheControlDirectiveTypedDict,
)
from .anthropicfiledocumentsource import (
AnthropicFileDocumentSource,
AnthropicFileDocumentSourceTypedDict,
)
from .anthropicimageblockparam import (
AnthropicImageBlockParam,
AnthropicImageBlockParamTypedDict,
@@ -89,6 +93,7 @@ AnthropicDocumentBlockParamSourceUnionTypedDict = TypeAliasType(
Union[
SourceContentTypedDict,
AnthropicURLPdfSourceTypedDict,
AnthropicFileDocumentSourceTypedDict,
AnthropicBase64PdfSourceTypedDict,
AnthropicPlainTextSourceTypedDict,
],
@@ -101,6 +106,7 @@ AnthropicDocumentBlockParamSourceUnion = Annotated[
Annotated[AnthropicPlainTextSource, Tag("text")],
Annotated[SourceContent, Tag("content")],
Annotated[AnthropicURLPdfSource, Tag("url")],
Annotated[AnthropicFileDocumentSource, Tag("file")],
],
Discriminator(lambda m: get_discriminator(m, "type", "type")),
]
@@ -0,0 +1,20 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import BaseModel
from typing import Literal
from typing_extensions import TypedDict
AnthropicFileDocumentSourceType = Literal["file",]
class AnthropicFileDocumentSourceTypedDict(TypedDict):
file_id: str
type: AnthropicFileDocumentSourceType
class AnthropicFileDocumentSource(BaseModel):
file_id: str
type: AnthropicFileDocumentSourceType
@@ -0,0 +1,17 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import BaseModel
from typing import Optional
from typing_extensions import NotRequired, TypedDict
class AnthropicIterationCacheCreationTypedDict(TypedDict):
ephemeral_1h_input_tokens: NotRequired[int]
ephemeral_5m_input_tokens: NotRequired[int]
class AnthropicIterationCacheCreation(BaseModel):
ephemeral_1h_input_tokens: Optional[int] = None
ephemeral_5m_input_tokens: Optional[int] = None
@@ -0,0 +1,83 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .anthropiciterationcachecreation import (
AnthropicIterationCacheCreation,
AnthropicIterationCacheCreationTypedDict,
)
from openrouter.types import (
BaseModel,
Nullable,
OptionalNullable,
UNSET,
UNSET_SENTINEL,
)
from pydantic import model_serializer
from typing import Literal, Optional
from typing_extensions import NotRequired, TypedDict
AnthropicMessageUsageIterationType = Literal["message",]
class AnthropicMessageUsageIterationTypedDict(TypedDict):
type: AnthropicMessageUsageIterationType
cache_creation: NotRequired[Nullable[AnthropicIterationCacheCreationTypedDict]]
cache_creation_input_tokens: NotRequired[int]
cache_read_input_tokens: NotRequired[int]
input_tokens: NotRequired[int]
output_tokens: NotRequired[int]
model: NotRequired[str]
class AnthropicMessageUsageIteration(BaseModel):
type: AnthropicMessageUsageIterationType
cache_creation: OptionalNullable[AnthropicIterationCacheCreation] = UNSET
cache_creation_input_tokens: Optional[int] = None
cache_read_input_tokens: Optional[int] = None
input_tokens: Optional[int] = None
output_tokens: Optional[int] = None
model: Optional[str] = None
@model_serializer(mode="wrap")
def serialize_model(self, handler):
optional_fields = [
"cache_creation",
"cache_creation_input_tokens",
"cache_read_input_tokens",
"input_tokens",
"output_tokens",
"model",
]
nullable_fields = ["cache_creation"]
null_default_fields = []
serialized = handler(self)
m = {}
for n, f in type(self).model_fields.items():
k = f.alias or n
val = serialized.get(k)
serialized.pop(k, None)
optional_nullable = k in optional_fields and k in nullable_fields
is_set = (
self.__pydantic_fields_set__.intersection({n})
or k in null_default_fields
) # pylint: disable=no-member
if val is not None and val != UNSET_SENTINEL:
m[k] = val
elif val != UNSET_SENTINEL and (
not k in optional_fields or (optional_nullable and is_set)
):
m[k] = val
return m
@@ -0,0 +1,14 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import UnrecognizedStr
from typing import Literal, Union
AnthropicSpeed = Union[
Literal[
"fast",
"standard",
],
UnrecognizedStr,
]
@@ -0,0 +1,76 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .anthropiciterationcachecreation import (
AnthropicIterationCacheCreation,
AnthropicIterationCacheCreationTypedDict,
)
from openrouter.types import (
BaseModel,
Nullable,
OptionalNullable,
UNSET,
UNSET_SENTINEL,
)
from pydantic import model_serializer
from typing import Optional
from typing_extensions import NotRequired, TypedDict
class AnthropicUnknownUsageIterationTypedDict(TypedDict):
type: str
cache_creation: NotRequired[Nullable[AnthropicIterationCacheCreationTypedDict]]
cache_creation_input_tokens: NotRequired[int]
cache_read_input_tokens: NotRequired[int]
input_tokens: NotRequired[int]
output_tokens: NotRequired[int]
class AnthropicUnknownUsageIteration(BaseModel):
type: str
cache_creation: OptionalNullable[AnthropicIterationCacheCreation] = UNSET
cache_creation_input_tokens: Optional[int] = None
cache_read_input_tokens: Optional[int] = None
input_tokens: Optional[int] = None
output_tokens: Optional[int] = None
@model_serializer(mode="wrap")
def serialize_model(self, handler):
optional_fields = [
"cache_creation",
"cache_creation_input_tokens",
"cache_read_input_tokens",
"input_tokens",
"output_tokens",
]
nullable_fields = ["cache_creation"]
null_default_fields = []
serialized = handler(self)
m = {}
for n, f in type(self).model_fields.items():
k = f.alias or n
val = serialized.get(k)
serialized.pop(k, None)
optional_nullable = k in optional_fields and k in nullable_fields
is_set = (
self.__pydantic_fields_set__.intersection({n})
or k in null_default_fields
) # pylint: disable=no-member
if val is not None and val != UNSET_SENTINEL:
m[k] = val
elif val != UNSET_SENTINEL and (
not k in optional_fields or (optional_nullable and is_set)
):
m[k] = val
return m
@@ -0,0 +1,43 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .anthropicadvisormessageusageiteration import (
AnthropicAdvisorMessageUsageIteration,
AnthropicAdvisorMessageUsageIterationTypedDict,
)
from .anthropiccompactionusageiteration import (
AnthropicCompactionUsageIteration,
AnthropicCompactionUsageIterationTypedDict,
)
from .anthropicmessageusageiteration import (
AnthropicMessageUsageIteration,
AnthropicMessageUsageIterationTypedDict,
)
from .anthropicunknownusageiteration import (
AnthropicUnknownUsageIteration,
AnthropicUnknownUsageIterationTypedDict,
)
from typing import Union
from typing_extensions import TypeAliasType
AnthropicUsageIterationTypedDict = TypeAliasType(
"AnthropicUsageIterationTypedDict",
Union[
AnthropicCompactionUsageIterationTypedDict,
AnthropicUnknownUsageIterationTypedDict,
AnthropicMessageUsageIterationTypedDict,
AnthropicAdvisorMessageUsageIterationTypedDict,
],
)
AnthropicUsageIteration = TypeAliasType(
"AnthropicUsageIteration",
Union[
AnthropicCompactionUsageIteration,
AnthropicUnknownUsageIteration,
AnthropicMessageUsageIteration,
AnthropicAdvisorMessageUsageIteration,
],
)
+40
View File
@@ -0,0 +1,40 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import UnrecognizedStr
from typing import Literal, Union
APIErrorType = Union[
Literal[
"context_length_exceeded",
"max_tokens_exceeded",
"token_limit_exceeded",
"string_too_long",
"authentication",
"permission_denied",
"payment_required",
"rate_limit_exceeded",
"provider_overloaded",
"provider_unavailable",
"invalid_request",
"invalid_prompt",
"not_found",
"precondition_failed",
"payload_too_large",
"unprocessable",
"content_policy_violation",
"refusal",
"invalid_image",
"image_too_large",
"image_too_small",
"unsupported_image_format",
"image_not_found",
"image_download_failed",
"server",
"timeout",
"unmapped",
],
UnrecognizedStr,
]
r"""Canonical OpenRouter error type, stable across all API formats"""
@@ -0,0 +1,21 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import BaseModel
from typing import Literal
from typing_extensions import TypedDict
BooleanCapabilityType = Literal["boolean",]
class BooleanCapabilityTypedDict(TypedDict):
r"""A supported-or-not flag. Present means the parameter is accepted."""
type: BooleanCapabilityType
class BooleanCapability(BaseModel):
r"""A supported-or-not flag. Present means the parameter is accepted."""
type: BooleanCapabilityType
@@ -43,8 +43,10 @@ BYOKProviderSlug = Union[
"google-ai-studio",
"google-vertex",
"groq",
"heygen",
"inception",
"inceptron",
"inferact-vllm",
"inference-net",
"infermatic",
"inflection",
@@ -72,9 +74,11 @@ BYOKProviderSlug = Union[
"perplexity",
"phala",
"poolside",
"quiver",
"recraft",
"reka",
"relace",
"sakana-ai",
"sambanova",
"seed",
"siliconflow",
@@ -82,6 +86,7 @@ BYOKProviderSlug = Union[
"stepfun",
"streamlake",
"switchpoint",
"tenstorrent",
"together",
"upstage",
"venice",
@@ -0,0 +1,30 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .booleancapability import BooleanCapability, BooleanCapabilityTypedDict
from .enumcapability import EnumCapability, EnumCapabilityTypedDict
from .rangecapability import RangeCapability, RangeCapabilityTypedDict
from openrouter.utils import get_discriminator
from pydantic import Discriminator, Tag
from typing import Union
from typing_extensions import Annotated, TypeAliasType
CapabilityDescriptorTypedDict = TypeAliasType(
"CapabilityDescriptorTypedDict",
Union[
BooleanCapabilityTypedDict, EnumCapabilityTypedDict, RangeCapabilityTypedDict
],
)
r"""A typed descriptor for one supported request parameter."""
CapabilityDescriptor = Annotated[
Union[
Annotated[EnumCapability, Tag("enum")],
Annotated[RangeCapability, Tag("range")],
Annotated[BooleanCapability, Tag("boolean")],
],
Discriminator(lambda m: get_discriminator(m, "type", "type")),
]
r"""A typed descriptor for one supported request parameter."""
@@ -27,6 +27,10 @@ from .openrouterwebsearchservertool import (
OpenRouterWebSearchServerTool,
OpenRouterWebSearchServerToolTypedDict,
)
from .subagentservertool_openrouter import (
SubagentServerToolOpenRouter,
SubagentServerToolOpenRouterTypedDict,
)
from .webfetchservertool import WebFetchServerTool, WebFetchServerToolTypedDict
from openrouter.types import (
BaseModel,
@@ -129,6 +133,7 @@ ChatFunctionToolTypedDict = TypeAliasType(
DatetimeServerToolTypedDict,
ImageGenerationServerToolOpenRouterTypedDict,
ChatSearchModelsServerToolTypedDict,
SubagentServerToolOpenRouterTypedDict,
WebFetchServerToolTypedDict,
OpenRouterWebSearchServerToolTypedDict,
ChatFunctionToolFunctionTypedDict,
@@ -150,6 +155,7 @@ ChatFunctionTool = Annotated[
Annotated[
ChatSearchModelsServerTool, Tag("openrouter:experimental__search_models")
],
Annotated[SubagentServerToolOpenRouter, Tag("openrouter:subagent")],
Annotated[WebFetchServerTool, Tag("openrouter:web_fetch")],
Annotated[OpenRouterWebSearchServerTool, Tag("openrouter:web_search")],
Annotated[ChatWebSearchShorthand, Tag("web_search")],
+54
View File
@@ -106,6 +106,7 @@ ChatRequestPlugin = Annotated[
ChatRequestEffort = Union[
Literal[
"max",
"xhigh",
"high",
"medium",
@@ -170,6 +171,21 @@ class ChatRequestReasoning(BaseModel):
return m
ChatRequestReasoningEffort = Union[
Literal[
"max",
"xhigh",
"high",
"medium",
"low",
"minimal",
"none",
],
UnrecognizedStr,
]
r"""Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ."""
ResponseFormatTypedDict = TypeAliasType(
"ResponseFormatTypedDict",
Union[
@@ -240,6 +256,8 @@ class ChatRequestTypedDict(TypedDict):
r"""Maximum tokens (deprecated, use max_completion_tokens). Note: some providers enforce a minimum of 16."""
metadata: NotRequired[Dict[str, str]]
r"""Key-value pairs for additional object information (max 16 pairs, 64 char keys, 512 char values)"""
min_p: NotRequired[Nullable[float]]
r"""Minimum probability threshold relative to the most likely token. Tokens with probability below min_p * (probability of top token) are filtered out. Not all providers support this parameter."""
modalities: NotRequired[List[Modality]]
r"""Output modalities for the response. Supported values are \"text\", \"image\", and \"audio\"."""
model: NotRequired[str]
@@ -256,6 +274,10 @@ class ChatRequestTypedDict(TypedDict):
r"""When multiple model providers are available, optionally indicate your routing preference."""
reasoning: NotRequired[ChatRequestReasoningTypedDict]
r"""Configuration options for reasoning models"""
reasoning_effort: NotRequired[Nullable[ChatRequestReasoningEffort]]
r"""Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ."""
repetition_penalty: NotRequired[Nullable[float]]
r"""Penalizes tokens based on how much they have already appeared in the text. A value of 1.0 means no penalty. Values above 1.0 penalize repeated tokens more strongly. Not all providers support this parameter."""
response_format: NotRequired[ResponseFormatTypedDict]
r"""Response format configuration"""
seed: NotRequired[Nullable[int]]
@@ -278,6 +300,10 @@ class ChatRequestTypedDict(TypedDict):
r"""Tool choice configuration"""
tools: NotRequired[List[ChatFunctionToolTypedDict]]
r"""Available tools for function calling"""
top_a: NotRequired[Nullable[float]]
r"""Consider only tokens with \"sufficiently high\" probabilities based on the probability of the most likely token. Not all providers support this parameter."""
top_k: NotRequired[Nullable[int]]
r"""Limits the model to choose from the top K most likely tokens at each step. A value of 1 means the model will always pick the most likely next token. Not all providers support this parameter."""
top_logprobs: NotRequired[Nullable[int]]
r"""Number of top log probabilities to return (0-20)"""
top_p: NotRequired[Nullable[float]]
@@ -321,6 +347,9 @@ class ChatRequest(BaseModel):
metadata: Optional[Dict[str, str]] = None
r"""Key-value pairs for additional object information (max 16 pairs, 64 char keys, 512 char values)"""
min_p: OptionalNullable[float] = UNSET
r"""Minimum probability threshold relative to the most likely token. Tokens with probability below min_p * (probability of top token) are filtered out. Not all providers support this parameter."""
modalities: Optional[
List[Annotated[Modality, PlainValidator(validate_open_enum(False))]]
] = None
@@ -347,6 +376,15 @@ class ChatRequest(BaseModel):
reasoning: Optional[ChatRequestReasoning] = None
r"""Configuration options for reasoning models"""
reasoning_effort: Annotated[
OptionalNullable[ChatRequestReasoningEffort],
PlainValidator(validate_open_enum(False)),
] = UNSET
r"""Shorthand for setting reasoning effort. Equivalent to setting reasoning.effort. Cannot be used simultaneously with reasoning.effort if they differ."""
repetition_penalty: OptionalNullable[float] = UNSET
r"""Penalizes tokens based on how much they have already appeared in the text. A value of 1.0 means no penalty. Values above 1.0 penalize repeated tokens more strongly. Not all providers support this parameter."""
response_format: Optional[ResponseFormat] = None
r"""Response format configuration"""
@@ -383,6 +421,12 @@ class ChatRequest(BaseModel):
tools: Optional[List[ChatFunctionTool]] = None
r"""Available tools for function calling"""
top_a: OptionalNullable[float] = UNSET
r"""Consider only tokens with \"sufficiently high\" probabilities based on the probability of the most likely token. Not all providers support this parameter."""
top_k: OptionalNullable[int] = UNSET
r"""Limits the model to choose from the top K most likely tokens at each step. A value of 1 means the model will always pick the most likely next token. Not all providers support this parameter."""
top_logprobs: OptionalNullable[int] = UNSET
r"""Number of top log probabilities to return (0-20)"""
@@ -407,6 +451,7 @@ class ChatRequest(BaseModel):
"max_completion_tokens",
"max_tokens",
"metadata",
"min_p",
"modalities",
"model",
"models",
@@ -415,6 +460,8 @@ class ChatRequest(BaseModel):
"presence_penalty",
"provider",
"reasoning",
"reasoning_effort",
"repetition_penalty",
"response_format",
"seed",
"service_tier",
@@ -426,6 +473,8 @@ class ChatRequest(BaseModel):
"temperature",
"tool_choice",
"tools",
"top_a",
"top_k",
"top_logprobs",
"top_p",
"trace",
@@ -437,14 +486,19 @@ class ChatRequest(BaseModel):
"logprobs",
"max_completion_tokens",
"max_tokens",
"min_p",
"parallel_tool_calls",
"presence_penalty",
"provider",
"reasoning_effort",
"repetition_penalty",
"seed",
"service_tier",
"stop",
"stream_options",
"temperature",
"top_a",
"top_k",
"top_logprobs",
"top_p",
]
+32 -5
View File
@@ -1,6 +1,7 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .apierrortype import APIErrorType
from .chatstreamchoice import ChatStreamChoice, ChatStreamChoiceTypedDict
from .chatusage import ChatUsage, ChatUsageTypedDict
from .openroutermetadata import OpenRouterMetadata, OpenRouterMetadataTypedDict
@@ -11,21 +12,44 @@ from openrouter.types import (
UNSET,
UNSET_SENTINEL,
)
from openrouter.utils import validate_open_enum
from pydantic import model_serializer
from pydantic.functional_validators import PlainValidator
from typing import List, Literal, Optional
from typing_extensions import NotRequired, TypedDict
from typing_extensions import Annotated, NotRequired, TypedDict
class ErrorTypedDict(TypedDict):
class ChatStreamChunkMetadataTypedDict(TypedDict):
r"""Structured error metadata"""
error_type: APIErrorType
r"""Canonical OpenRouter error type, stable across all API formats"""
provider_code: NotRequired[str]
r"""Upstream provider-specific error code, when available"""
class ChatStreamChunkMetadata(BaseModel):
r"""Structured error metadata"""
error_type: Annotated[APIErrorType, PlainValidator(validate_open_enum(False))]
r"""Canonical OpenRouter error type, stable across all API formats"""
provider_code: Optional[str] = None
r"""Upstream provider-specific error code, when available"""
class ChatStreamChunkErrorTypedDict(TypedDict):
r"""Error information"""
code: int
r"""Error code"""
message: str
r"""Error message"""
metadata: NotRequired[ChatStreamChunkMetadataTypedDict]
r"""Structured error metadata"""
class Error(BaseModel):
class ChatStreamChunkError(BaseModel):
r"""Error information"""
code: int
@@ -34,6 +58,9 @@ class Error(BaseModel):
message: str
r"""Error message"""
metadata: Optional[ChatStreamChunkMetadata] = None
r"""Structured error metadata"""
ChatStreamChunkObject = Literal["chat.completion.chunk",]
@@ -50,7 +77,7 @@ class ChatStreamChunkTypedDict(TypedDict):
model: str
r"""Model used for completion"""
object: ChatStreamChunkObject
error: NotRequired[ErrorTypedDict]
error: NotRequired[ChatStreamChunkErrorTypedDict]
r"""Error information"""
openrouter_metadata: NotRequired[OpenRouterMetadataTypedDict]
service_tier: NotRequired[Nullable[str]]
@@ -78,7 +105,7 @@ class ChatStreamChunk(BaseModel):
object: ChatStreamChunkObject
error: Optional[Error] = None
error: Optional[ChatStreamChunkError] = None
r"""Error information"""
openrouter_metadata: Optional[OpenRouterMetadata] = None
+123 -50
View File
@@ -14,7 +14,7 @@ from typing import Optional
from typing_extensions import NotRequired, TypedDict
class CompletionTokensDetailsTypedDict(TypedDict):
class ChatUsageCompletionTokensDetailsTypedDict(TypedDict):
r"""Detailed completion token usage"""
accepted_prediction_tokens: NotRequired[Nullable[int]]
@@ -27,7 +27,7 @@ class CompletionTokensDetailsTypedDict(TypedDict):
r"""Rejected prediction tokens"""
class CompletionTokensDetails(BaseModel):
class ChatUsageCompletionTokensDetails(BaseModel):
r"""Detailed completion token usage"""
accepted_prediction_tokens: OptionalNullable[int] = UNSET
@@ -83,7 +83,7 @@ class CompletionTokensDetails(BaseModel):
return m
class PromptTokensDetailsTypedDict(TypedDict):
class ChatUsagePromptTokensDetailsTypedDict(TypedDict):
r"""Detailed prompt token usage"""
audio_tokens: NotRequired[int]
@@ -96,7 +96,7 @@ class PromptTokensDetailsTypedDict(TypedDict):
r"""Video input tokens"""
class PromptTokensDetails(BaseModel):
class ChatUsagePromptTokensDetails(BaseModel):
r"""Detailed prompt token usage"""
audio_tokens: Optional[int] = None
@@ -112,68 +112,141 @@ class PromptTokensDetails(BaseModel):
r"""Video input tokens"""
class ChatUsageTypedDict(TypedDict):
r"""Token usage statistics"""
class ServerToolUseDetailsTypedDict(TypedDict):
r"""Usage for server-side tool execution (e.g., web search)"""
completion_tokens: int
r"""Number of tokens in the completion"""
prompt_tokens: int
r"""Number of tokens in the prompt"""
total_tokens: int
r"""Total number of tokens"""
completion_tokens_details: NotRequired[Nullable[CompletionTokensDetailsTypedDict]]
r"""Detailed completion token usage"""
cost: NotRequired[Nullable[float]]
r"""Cost of the completion"""
cost_details: NotRequired[Nullable[CostDetailsTypedDict]]
r"""Breakdown of upstream inference costs"""
is_byok: NotRequired[bool]
r"""Whether a request was made using a Bring Your Own Key configuration"""
prompt_tokens_details: NotRequired[Nullable[PromptTokensDetailsTypedDict]]
r"""Detailed prompt token usage"""
tool_calls_executed: NotRequired[Nullable[int]]
r"""Number of OpenRouter server tool calls that executed and produced a result"""
tool_calls_requested: NotRequired[Nullable[int]]
r"""Total number of OpenRouter server-orchestrated tool calls the model requested, across all tool types. Provider-native tools (e.g. native web search) are not counted here."""
web_search_requests: NotRequired[Nullable[int]]
r"""Number of web searches performed by server-side tools. For server-orchestrated tool calls a web search is also counted in tool_calls_requested; provider-native web search may report web_search_requests only. Do not sum the two."""
class ChatUsage(BaseModel):
r"""Token usage statistics"""
class ServerToolUseDetails(BaseModel):
r"""Usage for server-side tool execution (e.g., web search)"""
completion_tokens: int
r"""Number of tokens in the completion"""
tool_calls_executed: OptionalNullable[int] = UNSET
r"""Number of OpenRouter server tool calls that executed and produced a result"""
prompt_tokens: int
r"""Number of tokens in the prompt"""
tool_calls_requested: OptionalNullable[int] = UNSET
r"""Total number of OpenRouter server-orchestrated tool calls the model requested, across all tool types. Provider-native tools (e.g. native web search) are not counted here."""
total_tokens: int
r"""Total number of tokens"""
completion_tokens_details: OptionalNullable[CompletionTokensDetails] = UNSET
r"""Detailed completion token usage"""
cost: OptionalNullable[float] = UNSET
r"""Cost of the completion"""
cost_details: OptionalNullable[CostDetails] = UNSET
r"""Breakdown of upstream inference costs"""
is_byok: Optional[bool] = None
r"""Whether a request was made using a Bring Your Own Key configuration"""
prompt_tokens_details: OptionalNullable[PromptTokensDetails] = UNSET
r"""Detailed prompt token usage"""
web_search_requests: OptionalNullable[int] = UNSET
r"""Number of web searches performed by server-side tools. For server-orchestrated tool calls a web search is also counted in tool_calls_requested; provider-native web search may report web_search_requests only. Do not sum the two."""
@model_serializer(mode="wrap")
def serialize_model(self, handler):
optional_fields = [
"completion_tokens_details",
"cost",
"cost_details",
"is_byok",
"prompt_tokens_details",
"tool_calls_executed",
"tool_calls_requested",
"web_search_requests",
]
nullable_fields = [
"tool_calls_executed",
"tool_calls_requested",
"web_search_requests",
]
null_default_fields = []
serialized = handler(self)
m = {}
for n, f in type(self).model_fields.items():
k = f.alias or n
val = serialized.get(k)
serialized.pop(k, None)
optional_nullable = k in optional_fields and k in nullable_fields
is_set = (
self.__pydantic_fields_set__.intersection({n})
or k in null_default_fields
) # pylint: disable=no-member
if val is not None and val != UNSET_SENTINEL:
m[k] = val
elif val != UNSET_SENTINEL and (
not k in optional_fields or (optional_nullable and is_set)
):
m[k] = val
return m
class ChatUsageTypedDict(TypedDict):
r"""Token usage statistics"""
completion_tokens: int
r"""Number of tokens in the completion"""
prompt_tokens: int
r"""Number of tokens in the prompt"""
total_tokens: int
r"""Total number of tokens"""
completion_tokens_details: NotRequired[
Nullable[ChatUsageCompletionTokensDetailsTypedDict]
]
r"""Detailed completion token usage"""
cost: NotRequired[Nullable[float]]
r"""Cost of the completion"""
cost_details: NotRequired[Nullable[CostDetailsTypedDict]]
r"""Breakdown of upstream inference costs"""
is_byok: NotRequired[bool]
r"""Whether a request was made using a Bring Your Own Key configuration"""
prompt_tokens_details: NotRequired[Nullable[ChatUsagePromptTokensDetailsTypedDict]]
r"""Detailed prompt token usage"""
server_tool_use_details: NotRequired[Nullable[ServerToolUseDetailsTypedDict]]
r"""Usage for server-side tool execution (e.g., web search)"""
class ChatUsage(BaseModel):
r"""Token usage statistics"""
completion_tokens: int
r"""Number of tokens in the completion"""
prompt_tokens: int
r"""Number of tokens in the prompt"""
total_tokens: int
r"""Total number of tokens"""
completion_tokens_details: OptionalNullable[ChatUsageCompletionTokensDetails] = (
UNSET
)
r"""Detailed completion token usage"""
cost: OptionalNullable[float] = UNSET
r"""Cost of the completion"""
cost_details: OptionalNullable[CostDetails] = UNSET
r"""Breakdown of upstream inference costs"""
is_byok: Optional[bool] = None
r"""Whether a request was made using a Bring Your Own Key configuration"""
prompt_tokens_details: OptionalNullable[ChatUsagePromptTokensDetails] = UNSET
r"""Detailed prompt token usage"""
server_tool_use_details: OptionalNullable[ServerToolUseDetails] = UNSET
r"""Usage for server-side tool execution (e.g., web search)"""
@model_serializer(mode="wrap")
def serialize_model(self, handler):
optional_fields = [
"completion_tokens_details",
"cost",
"cost_details",
"is_byok",
"prompt_tokens_details",
"server_tool_use_details",
]
nullable_fields = [
"completion_tokens_details",
"cost",
"cost_details",
"prompt_tokens_details",
"server_tool_use_details",
]
null_default_fields = []
@@ -31,20 +31,20 @@ class ChatWebSearchShorthandTypedDict(TypedDict):
type: ChatWebSearchShorthandType
allowed_domains: NotRequired[List[str]]
r"""Limit search results to these domains. Supported by Exa, Firecrawl, Parallel, and most native providers (Anthropic, OpenAI, xAI). Not supported with Perplexity. Cannot be used with excluded_domains."""
r"""Limit search results to these domains. Supported by Exa, Firecrawl, Parallel, Perplexity, and most native providers (Anthropic, OpenAI, xAI). Cannot be used with excluded_domains."""
engine: NotRequired[WebSearchEngineEnum]
r"""Which search engine to use. \"auto\" (default) uses native if the provider supports it, otherwise Exa. \"native\" forces the provider's built-in search. \"exa\" forces the Exa search API. \"firecrawl\" uses Firecrawl (requires BYOK). \"parallel\" uses the Parallel search API."""
r"""Which search engine to use. \"auto\" (default) uses native if the provider supports it, otherwise Exa. \"native\" forces the provider's built-in search. \"exa\" forces the Exa search API. \"firecrawl\" uses Firecrawl (requires BYOK). \"parallel\" uses the Parallel search API. \"perplexity\" uses the Perplexity Search API (raw ranked results)."""
excluded_domains: NotRequired[List[str]]
r"""Exclude search results from these domains. Supported by Exa, Firecrawl, Parallel, Anthropic, and xAI. Not supported with OpenAI (silently ignored) or Perplexity. Cannot be used with allowed_domains."""
r"""Exclude search results from these domains. Supported by Exa, Firecrawl, Parallel, Perplexity, Anthropic, and xAI. Not supported with OpenAI (silently ignored). Cannot be used with allowed_domains."""
max_characters: NotRequired[int]
r"""Exact maximum number of characters of content per search result. Applies to the Exa and Parallel engines; ignored with native provider search and Firecrawl. For Exa, caps highlight content per result. For Parallel, caps excerpt content per result (default 1,500 when omitted). When both `max_characters` and `search_context_size` are set, `max_characters` takes precedence for both engines. When omitted, falls back to `search_context_size` mapping (Exa) or engine defaults (Parallel)."""
r"""Exact maximum number of characters of content per search result. Applies to the Exa, Parallel, and Perplexity engines; ignored with native provider search and Firecrawl. For Exa, caps highlight content per result. For Parallel, caps excerpt content per result (default 1,500 when omitted). For Perplexity, maps to the native `max_tokens_per_page` parameter (converted from characters to tokens) and trims the response to the exact character cap. When both `max_characters` and `search_context_size` are set, `max_characters` takes precedence. When omitted, falls back to `search_context_size` mapping (Exa) or engine defaults (Parallel, Perplexity)."""
max_results: NotRequired[int]
r"""Maximum number of search results to return per search call. Defaults to 5. Applies to Exa, Firecrawl, and Parallel engines; ignored with native provider search."""
r"""Maximum number of search results to return per search call. Defaults to 5. Applies to Exa, Firecrawl, Parallel, and Perplexity engines; ignored with native provider search. Perplexity supports a maximum of 20; values above 20 are clamped."""
max_total_results: NotRequired[int]
r"""Maximum total number of search results across all search calls in a single request. Once this limit is reached, the tool will stop returning new results. Useful for controlling cost and context size in agentic loops. Defaults to 50 when not specified."""
parameters: NotRequired[WebSearchConfigTypedDict]
search_context_size: NotRequired[SearchQualityLevel]
r"""How much context to retrieve per result. Applies to Exa and Parallel engines; ignored with native provider search and Firecrawl. For Exa, pins a fixed per-result character cap (low=5,000, medium=15,000, high=30,000); when omitted, Exa picks an adaptive size per query and document (typically ~2,0004,000 characters per result). For Parallel, controls the total characters across all results; when omitted, Parallel uses its own default size. Overridden by `max_characters` when both are set."""
r"""How much context to retrieve per result. Applies to Exa, Parallel, and Perplexity engines; ignored with native provider search and Firecrawl. For Exa, pins a fixed per-result character cap (low=5,000, medium=15,000, high=30,000); when omitted, Exa picks an adaptive size per query and document (typically ~2,0004,000 characters per result). For Parallel, controls the total characters across all results; when omitted, Parallel uses its own default size. For Perplexity, maps directly to the Search API's native search_context_size parameter. Overridden by `max_characters` when both are set."""
user_location: NotRequired[WebSearchUserLocationServerToolTypedDict]
r"""Approximate user location for location-biased results."""
@@ -57,21 +57,21 @@ class ChatWebSearchShorthand(BaseModel):
]
allowed_domains: Optional[List[str]] = None
r"""Limit search results to these domains. Supported by Exa, Firecrawl, Parallel, and most native providers (Anthropic, OpenAI, xAI). Not supported with Perplexity. Cannot be used with excluded_domains."""
r"""Limit search results to these domains. Supported by Exa, Firecrawl, Parallel, Perplexity, and most native providers (Anthropic, OpenAI, xAI). Cannot be used with excluded_domains."""
engine: Annotated[
Optional[WebSearchEngineEnum], PlainValidator(validate_open_enum(False))
] = None
r"""Which search engine to use. \"auto\" (default) uses native if the provider supports it, otherwise Exa. \"native\" forces the provider's built-in search. \"exa\" forces the Exa search API. \"firecrawl\" uses Firecrawl (requires BYOK). \"parallel\" uses the Parallel search API."""
r"""Which search engine to use. \"auto\" (default) uses native if the provider supports it, otherwise Exa. \"native\" forces the provider's built-in search. \"exa\" forces the Exa search API. \"firecrawl\" uses Firecrawl (requires BYOK). \"parallel\" uses the Parallel search API. \"perplexity\" uses the Perplexity Search API (raw ranked results)."""
excluded_domains: Optional[List[str]] = None
r"""Exclude search results from these domains. Supported by Exa, Firecrawl, Parallel, Anthropic, and xAI. Not supported with OpenAI (silently ignored) or Perplexity. Cannot be used with allowed_domains."""
r"""Exclude search results from these domains. Supported by Exa, Firecrawl, Parallel, Perplexity, Anthropic, and xAI. Not supported with OpenAI (silently ignored). Cannot be used with allowed_domains."""
max_characters: Optional[int] = None
r"""Exact maximum number of characters of content per search result. Applies to the Exa and Parallel engines; ignored with native provider search and Firecrawl. For Exa, caps highlight content per result. For Parallel, caps excerpt content per result (default 1,500 when omitted). When both `max_characters` and `search_context_size` are set, `max_characters` takes precedence for both engines. When omitted, falls back to `search_context_size` mapping (Exa) or engine defaults (Parallel)."""
r"""Exact maximum number of characters of content per search result. Applies to the Exa, Parallel, and Perplexity engines; ignored with native provider search and Firecrawl. For Exa, caps highlight content per result. For Parallel, caps excerpt content per result (default 1,500 when omitted). For Perplexity, maps to the native `max_tokens_per_page` parameter (converted from characters to tokens) and trims the response to the exact character cap. When both `max_characters` and `search_context_size` are set, `max_characters` takes precedence. When omitted, falls back to `search_context_size` mapping (Exa) or engine defaults (Parallel, Perplexity)."""
max_results: Optional[int] = None
r"""Maximum number of search results to return per search call. Defaults to 5. Applies to Exa, Firecrawl, and Parallel engines; ignored with native provider search."""
r"""Maximum number of search results to return per search call. Defaults to 5. Applies to Exa, Firecrawl, Parallel, and Perplexity engines; ignored with native provider search. Perplexity supports a maximum of 20; values above 20 are clamped."""
max_total_results: Optional[int] = None
r"""Maximum total number of search results across all search calls in a single request. Once this limit is reached, the tool will stop returning new results. Useful for controlling cost and context size in agentic loops. Defaults to 50 when not specified."""
@@ -81,7 +81,7 @@ class ChatWebSearchShorthand(BaseModel):
search_context_size: Annotated[
Optional[SearchQualityLevel], PlainValidator(validate_open_enum(False))
] = None
r"""How much context to retrieve per result. Applies to Exa and Parallel engines; ignored with native provider search and Firecrawl. For Exa, pins a fixed per-result character cap (low=5,000, medium=15,000, high=30,000); when omitted, Exa picks an adaptive size per query and document (typically ~2,0004,000 characters per result). For Parallel, controls the total characters across all results; when omitted, Parallel uses its own default size. Overridden by `max_characters` when both are set."""
r"""How much context to retrieve per result. Applies to Exa, Parallel, and Perplexity engines; ignored with native provider search and Firecrawl. For Exa, pins a fixed per-result character cap (low=5,000, medium=15,000, high=30,000); when omitted, Exa picks an adaptive size per query and document (typically ~2,0004,000 characters per result). For Parallel, controls the total characters across all results; when omitted, Parallel uses its own default size. For Perplexity, maps directly to the Search API's native search_context_size parameter. Overridden by `max_characters` when both are set."""
user_location: Optional[WebSearchUserLocationServerTool] = None
r"""Approximate user location for location-biased results."""
@@ -0,0 +1,39 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import BaseModel
from typing_extensions import TypedDict
class DABenchmarkEntryTypedDict(TypedDict):
r"""A single Design Arena benchmark entry for a specific arena+category"""
arena: str
r"""Arena type (e.g. models, builders, agents)"""
category: str
r"""Category within the arena (e.g. website, gamedev, uicomponent)"""
elo: float
r"""ELO rating from head-to-head arena battles"""
rank: int
r"""Rank position within this arena+category among models available on OpenRouter (1 = highest ELO)"""
win_rate: float
r"""Win rate percentage in arena battles"""
class DABenchmarkEntry(BaseModel):
r"""A single Design Arena benchmark entry for a specific arena+category"""
arena: str
r"""Arena type (e.g. models, builders, agents)"""
category: str
r"""Category within the arena (e.g. website, gamedev, uicomponent)"""
elo: float
r"""ELO rating from head-to-head arena battles"""
rank: int
r"""Rank position within this arena+category among models available on OpenRouter (1 = highest ELO)"""
win_rate: float
r"""Win rate percentage in arena battles"""
+65
View File
@@ -0,0 +1,65 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import BaseModel, Nullable, UnrecognizedStr
from openrouter.utils import validate_open_enum
from pydantic.functional_validators import PlainValidator
from typing import Any, Dict, Literal, Optional, Union
from typing_extensions import Annotated, NotRequired, TypedDict
Event = Union[
Literal[
"adapter_request",
"upstream_headers_received",
"first_token_received",
"upstream_body_ended",
],
UnrecognizedStr,
]
class TimingsTypedDict(TypedDict):
epoch_ms: int
event: Event
start_ms: int
class Timings(BaseModel):
epoch_ms: int
event: Annotated[Event, PlainValidator(validate_open_enum(False))]
start_ms: int
class DebugTypedDict(TypedDict):
echo_upstream_body: NotRequired[Dict[str, Nullable[Any]]]
timings: NotRequired[TimingsTypedDict]
class Debug(BaseModel):
echo_upstream_body: Optional[Dict[str, Nullable[Any]]] = None
timings: Optional[Timings] = None
DebugEventType = Literal["response.debug",]
class DebugEventTypedDict(TypedDict):
r"""Debug event emitted when debug.echo_upstream_body is true. Contains the transformed upstream request body or timing milestones."""
debug: DebugTypedDict
sequence_number: int
type: DebugEventType
class DebugEvent(BaseModel):
r"""Debug event emitted when debug.echo_upstream_body is true. Contains the transformed upstream request body or timing milestones."""
debug: Debug
sequence_number: int
type: DebugEventType
@@ -0,0 +1,22 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import BaseModel
from openrouter.utils import validate_const
import pydantic
from pydantic.functional_validators import AfterValidator
from typing import Literal
from typing_extensions import Annotated, TypedDict
class DeleteWorkspaceBudgetResponseTypedDict(TypedDict):
deleted: Literal[True]
r"""Confirmation that the budget was deleted (or did not exist)"""
class DeleteWorkspaceBudgetResponse(BaseModel):
DELETED: Annotated[
Annotated[Literal[True], AfterValidator(validate_const(True))],
pydantic.Field(alias="deleted"),
] = True
r"""Confirmation that the budget was deleted (or did not exist)"""
@@ -0,0 +1,24 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import BaseModel
from typing import List, Literal
from typing_extensions import TypedDict
EnumCapabilityType = Literal["enum",]
class EnumCapabilityTypedDict(TypedDict):
r"""A parameter that accepts one of a discrete set of string values."""
type: EnumCapabilityType
values: List[str]
class EnumCapability(BaseModel):
r"""A parameter that accepts one of a discrete set of string values."""
type: EnumCapabilityType
values: List[str]
@@ -0,0 +1,24 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import BaseModel
from typing import Literal
from typing_extensions import TypedDict
FileDeleteResponseType = Literal["file_deleted",]
class FileDeleteResponseTypedDict(TypedDict):
r"""Confirmation that a file was deleted."""
id: str
type: FileDeleteResponseType
class FileDeleteResponse(BaseModel):
r"""Confirmation that a file was deleted."""
id: str
type: FileDeleteResponseType
@@ -0,0 +1,64 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .filemetadata import FileMetadata, FileMetadataTypedDict
from openrouter.types import BaseModel, Nullable, UNSET_SENTINEL
from pydantic import model_serializer
from typing import List
from typing_extensions import TypedDict
class FileListResponseTypedDict(TypedDict):
r"""A page of files belonging to the requesting workspace."""
cursor: Nullable[str]
r"""Opaque cursor for the next page; null when there are no more results."""
data: List[FileMetadataTypedDict]
first_id: Nullable[str]
has_more: bool
last_id: Nullable[str]
class FileListResponse(BaseModel):
r"""A page of files belonging to the requesting workspace."""
cursor: Nullable[str]
r"""Opaque cursor for the next page; null when there are no more results."""
data: List[FileMetadata]
first_id: Nullable[str]
has_more: bool
last_id: Nullable[str]
@model_serializer(mode="wrap")
def serialize_model(self, handler):
optional_fields = []
nullable_fields = ["cursor", "first_id", "last_id"]
null_default_fields = []
serialized = handler(self)
m = {}
for n, f in type(self).model_fields.items():
k = f.alias or n
val = serialized.get(k)
serialized.pop(k, None)
optional_nullable = k in optional_fields and k in nullable_fields
is_set = (
self.__pydantic_fields_set__.intersection({n})
or k in null_default_fields
) # pylint: disable=no-member
if val is not None and val != UNSET_SENTINEL:
m[k] = val
elif val != UNSET_SENTINEL and (
not k in optional_fields or (optional_nullable and is_set)
):
m[k] = val
return m
+39
View File
@@ -0,0 +1,39 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import BaseModel
from typing import Literal
from typing_extensions import TypedDict
FileMetadataType = Literal["file",]
class FileMetadataTypedDict(TypedDict):
r"""Metadata describing a stored file."""
created_at: str
downloadable: bool
filename: str
id: str
mime_type: str
size_bytes: int
type: FileMetadataType
class FileMetadata(BaseModel):
r"""Metadata describing a stored file."""
created_at: str
downloadable: bool
filename: str
id: str
mime_type: str
size_bytes: int
type: FileMetadataType
+23 -3
View File
@@ -1,14 +1,27 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import BaseModel, Nullable
from typing import Any, Dict, List, Literal, Optional
from typing_extensions import NotRequired, TypedDict
from openrouter.types import BaseModel, Nullable, UnrecognizedStr
from openrouter.utils import validate_open_enum
from pydantic.functional_validators import PlainValidator
from typing import Any, Dict, List, Literal, Optional, Union
from typing_extensions import Annotated, NotRequired, TypedDict
FusionPluginID = Literal["fusion",]
PresetEnum = Union[
Literal[
"general-high",
"general-budget",
"general-fast",
],
UnrecognizedStr,
]
r"""A curated OpenRouter fusion preset (slugs follow `<task>-<tier>`, e.g. `general-high`). Expands server-side into the preset's analysis_models panel and judge model, so callers never name individual models. Explicitly provided `analysis_models` / `model` take precedence."""
class FusionPluginToolTypedDict(TypedDict):
type: str
r"""Server tool type identifier (e.g. \"openrouter:web_search\", \"openrouter:web_fetch\")."""
@@ -34,6 +47,8 @@ class FusionPluginTypedDict(TypedDict):
r"""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."""
model: NotRequired[str]
r"""Slug of the model that performs both the judge step (with web_search + web_fetch) and the final synthesis. When omitted, defaults to the first model in the Quality preset."""
preset: NotRequired[PresetEnum]
r"""A curated OpenRouter fusion preset (slugs follow `<task>-<tier>`, e.g. `general-high`). Expands server-side into the preset's analysis_models panel and judge model, so callers never name individual models. Explicitly provided `analysis_models` / `model` take precedence."""
tools: NotRequired[List[FusionPluginToolTypedDict]]
r"""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)."""
@@ -53,5 +68,10 @@ class FusionPlugin(BaseModel):
model: Optional[str] = None
r"""Slug of the model that performs both the judge step (with web_search + web_fetch) and the final synthesis. When omitted, defaults to the first model in the Quality preset."""
preset: Annotated[
Optional[PresetEnum], PlainValidator(validate_open_enum(False))
] = None
r"""A curated OpenRouter fusion preset (slugs follow `<task>-<tier>`, e.g. `general-high`). Expands server-side into the preset's analysis_models panel and judge model, so callers never name individual models. Explicitly provided `analysis_models` / `model` take precedence."""
tools: Optional[List[FusionPluginTool]] = None
r"""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)."""
@@ -1,6 +1,10 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .anthropiccachecontroldirective import (
AnthropicCacheControlDirective,
AnthropicCacheControlDirectiveTypedDict,
)
from openrouter.types import BaseModel, Nullable, UnrecognizedStr
from openrouter.utils import validate_open_enum
from pydantic.functional_validators import PlainValidator
@@ -10,6 +14,7 @@ from typing_extensions import Annotated, NotRequired, TypedDict
FusionServerToolConfigEffort = Union[
Literal[
"max",
"xhigh",
"high",
"medium",
@@ -64,6 +69,8 @@ class FusionServerToolConfigTypedDict(TypedDict):
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."""
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."""
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, the provider's default applies."""
max_tool_calls: NotRequired[int]
@@ -73,7 +80,7 @@ class FusionServerToolConfigTypedDict(TypedDict):
reasoning: NotRequired[FusionServerToolConfigReasoningTypedDict]
r"""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: NotRequired[float]
r"""Sampling temperature forwarded to panelist and judge inner calls. When omitted, the provider's default applies."""
r"""Temperature forwarded to panelist inner calls. The judge always runs at temperature 0 regardless of this value. When omitted, the provider's default applies."""
tools: NotRequired[List[FusionServerToolConfigToolTypedDict]]
r"""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)."""
@@ -84,6 +91,9 @@ class FusionServerToolConfig(BaseModel):
analysis_models: Optional[List[str]] = None
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
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."""
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, the provider's default applies."""
@@ -97,7 +107,7 @@ class FusionServerToolConfig(BaseModel):
r"""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] = None
r"""Sampling temperature forwarded to panelist and judge inner calls. When omitted, the provider's default applies."""
r"""Temperature forwarded to panelist inner calls. The judge always runs at temperature 0 regardless of this value. When omitted, the provider's default applies."""
tools: Optional[List[FusionServerToolConfigTool]] = None
r"""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)."""
+24
View File
@@ -0,0 +1,24 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import BaseModel
from typing_extensions import TypedDict
class FusionSourceTypedDict(TypedDict):
r"""A web page retrieved via web search during a fusion run."""
title: str
r"""Title of the retrieved web page."""
url: str
r"""URL of the web page a panel or the judge retrieved during the run."""
class FusionSource(BaseModel):
r"""A web page retrieved via web search during a fusion run."""
title: str
r"""Title of the retrieved web page."""
url: str
r"""URL of the web page a panel or the judge retrieved during the run."""
@@ -0,0 +1,81 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .capabilitydescriptor import CapabilityDescriptor, CapabilityDescriptorTypedDict
from .imagepricingentry import ImagePricingEntry, ImagePricingEntryTypedDict
from openrouter.types import BaseModel, Nullable, UNSET_SENTINEL
from pydantic import model_serializer
from typing import Dict, List
from typing_extensions import TypedDict
class ImageEndpointTypedDict(TypedDict):
r"""An endpoint that serves a given image model."""
allowed_passthrough_parameters: List[str]
r"""Provider-specific options accepted under provider.options[provider_slug]."""
pricing: List[ImagePricingEntryTypedDict]
r"""Billable pricing lines for this endpoint."""
provider_name: str
r"""Provider display name"""
provider_slug: str
r"""Provider slug"""
provider_tag: Nullable[str]
r"""Provider tag for request-side selection"""
supported_parameters: Dict[str, CapabilityDescriptorTypedDict]
supports_streaming: bool
r"""Whether this endpoint supports native SSE streaming (`stream: true` in the request)."""
class ImageEndpoint(BaseModel):
r"""An endpoint that serves a given image model."""
allowed_passthrough_parameters: List[str]
r"""Provider-specific options accepted under provider.options[provider_slug]."""
pricing: List[ImagePricingEntry]
r"""Billable pricing lines for this endpoint."""
provider_name: str
r"""Provider display name"""
provider_slug: str
r"""Provider slug"""
provider_tag: Nullable[str]
r"""Provider tag for request-side selection"""
supported_parameters: Dict[str, CapabilityDescriptor]
supports_streaming: bool
r"""Whether this endpoint supports native SSE streaming (`stream: true` in the request)."""
@model_serializer(mode="wrap")
def serialize_model(self, handler):
optional_fields = []
nullable_fields = ["provider_tag"]
null_default_fields = []
serialized = handler(self)
m = {}
for n, f in type(self).model_fields.items():
k = f.alias or n
val = serialized.get(k)
serialized.pop(k, None)
optional_nullable = k in optional_fields and k in nullable_fields
is_set = (
self.__pydantic_fields_set__.intersection({n})
or k in null_default_fields
) # pylint: disable=no-member
if val is not None and val != UNSET_SENTINEL:
m[k] = val
elif val != UNSET_SENTINEL and (
not k in optional_fields or (optional_nullable and is_set)
):
m[k] = val
return m
@@ -0,0 +1,45 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .imagegenerationusage import ImageGenerationUsage, ImageGenerationUsageTypedDict
from openrouter.types import BaseModel
from typing import Literal, Optional
from typing_extensions import NotRequired, TypedDict
ImageGenCompletedEventType = Literal["image_generation.completed",]
r"""The event type"""
class ImageGenCompletedEventTypedDict(TypedDict):
r"""Emitted when generation completes and the final image is available"""
b64_json: str
r"""Base64-encoded final image data"""
created: int
r"""Unix timestamp (seconds) when the image was generated"""
type: ImageGenCompletedEventType
r"""The event type"""
media_type: NotRequired[str]
r"""Media type (MIME type) of the image. Omitted when the output is a standard raster format (PNG). Present for non-raster outputs such as SVG (`image/svg+xml`)."""
usage: NotRequired[ImageGenerationUsageTypedDict]
r"""Token and cost usage for the image generation request, when available"""
class ImageGenCompletedEvent(BaseModel):
r"""Emitted when generation completes and the final image is available"""
b64_json: str
r"""Base64-encoded final image data"""
created: int
r"""Unix timestamp (seconds) when the image was generated"""
type: ImageGenCompletedEventType
r"""The event type"""
media_type: Optional[str] = None
r"""Media type (MIME type) of the image. Omitted when the output is a standard raster format (PNG). Present for non-raster outputs such as SVG (`image/svg+xml`)."""
usage: Optional[ImageGenerationUsage] = None
r"""Token and cost usage for the image generation request, when available"""
@@ -0,0 +1,610 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .contentpartimage import ContentPartImage, ContentPartImageTypedDict
from openrouter.types import BaseModel, Nullable, UnrecognizedStr
from openrouter.utils import validate_open_enum
import pydantic
from pydantic.functional_validators import PlainValidator
from typing import Any, Dict, List, Literal, Optional, Union
from typing_extensions import Annotated, NotRequired, TypedDict
ImageGenerationRequestAspectRatio = Union[
Literal[
"1:1",
"1:2",
"1:4",
"1:8",
"2:1",
"2:3",
"3:2",
"3:4",
"4:1",
"4:3",
"4:5",
"5:4",
"8:1",
"9:16",
"16:9",
"9:19.5",
"19.5:9",
"9:20",
"20:9",
"9:21",
"21:9",
"auto",
],
UnrecognizedStr,
]
r"""Normalized aspect ratio of the generated image. Providers clamp to their supported subset."""
ImageGenerationRequestBackground = Union[
Literal[
"auto",
"transparent",
"opaque",
],
UnrecognizedStr,
]
r"""Background treatment. `transparent` requires an output_format that supports alpha (png or webp)."""
ImageGenerationRequestOutputFormat = Union[
Literal[
"png",
"jpeg",
"webp",
"svg",
],
UnrecognizedStr,
]
r"""Encoding of the returned image bytes. Most models produce raster formats (png, jpeg, webp). SVG is supported by vectorization models (e.g. Quiver) — the SVG markup is UTF-8 base64-encoded in `b64_json`."""
class ImageGenerationRequestOptionsTypedDict(TypedDict):
r"""Provider-specific options keyed by provider slug. Only options for the matched provider are forwarded; the rest are ignored. Unrecognized keys are silently dropped."""
oneai: NotRequired[Dict[str, Nullable[Any]]]
ai21: NotRequired[Dict[str, Nullable[Any]]]
aion_labs: NotRequired[Dict[str, Nullable[Any]]]
akashml: NotRequired[Dict[str, Nullable[Any]]]
alibaba: NotRequired[Dict[str, Nullable[Any]]]
amazon_bedrock: NotRequired[Dict[str, Nullable[Any]]]
amazon_nova: NotRequired[Dict[str, Nullable[Any]]]
ambient: NotRequired[Dict[str, Nullable[Any]]]
anthropic: NotRequired[Dict[str, Nullable[Any]]]
anyscale: NotRequired[Dict[str, Nullable[Any]]]
arcee_ai: NotRequired[Dict[str, Nullable[Any]]]
atlas_cloud: NotRequired[Dict[str, Nullable[Any]]]
atoma: NotRequired[Dict[str, Nullable[Any]]]
avian: NotRequired[Dict[str, Nullable[Any]]]
azure: NotRequired[Dict[str, Nullable[Any]]]
baidu: NotRequired[Dict[str, Nullable[Any]]]
baseten: NotRequired[Dict[str, Nullable[Any]]]
black_forest_labs: NotRequired[Dict[str, Nullable[Any]]]
byteplus: NotRequired[Dict[str, Nullable[Any]]]
centml: NotRequired[Dict[str, Nullable[Any]]]
cerebras: NotRequired[Dict[str, Nullable[Any]]]
chutes: NotRequired[Dict[str, Nullable[Any]]]
cirrascale: NotRequired[Dict[str, Nullable[Any]]]
clarifai: NotRequired[Dict[str, Nullable[Any]]]
cloudflare: NotRequired[Dict[str, Nullable[Any]]]
cohere: NotRequired[Dict[str, Nullable[Any]]]
crofai: NotRequired[Dict[str, Nullable[Any]]]
crucible: NotRequired[Dict[str, Nullable[Any]]]
crusoe: NotRequired[Dict[str, Nullable[Any]]]
darkbloom: NotRequired[Dict[str, Nullable[Any]]]
decart: NotRequired[Dict[str, Nullable[Any]]]
deepinfra: NotRequired[Dict[str, Nullable[Any]]]
deepseek: NotRequired[Dict[str, Nullable[Any]]]
dekallm: NotRequired[Dict[str, Nullable[Any]]]
digitalocean: NotRequired[Dict[str, Nullable[Any]]]
enfer: NotRequired[Dict[str, Nullable[Any]]]
fake_provider: NotRequired[Dict[str, Nullable[Any]]]
featherless: NotRequired[Dict[str, Nullable[Any]]]
fireworks: NotRequired[Dict[str, Nullable[Any]]]
friendli: NotRequired[Dict[str, Nullable[Any]]]
gmicloud: NotRequired[Dict[str, Nullable[Any]]]
google_ai_studio: NotRequired[Dict[str, Nullable[Any]]]
google_vertex: NotRequired[Dict[str, Nullable[Any]]]
gopomelo: NotRequired[Dict[str, Nullable[Any]]]
groq: NotRequired[Dict[str, Nullable[Any]]]
heygen: NotRequired[Dict[str, Nullable[Any]]]
huggingface: NotRequired[Dict[str, Nullable[Any]]]
hyperbolic: NotRequired[Dict[str, Nullable[Any]]]
hyperbolic_quantized: NotRequired[Dict[str, Nullable[Any]]]
inception: NotRequired[Dict[str, Nullable[Any]]]
inceptron: NotRequired[Dict[str, Nullable[Any]]]
inferact_vllm: NotRequired[Dict[str, Nullable[Any]]]
inference_net: NotRequired[Dict[str, Nullable[Any]]]
infermatic: NotRequired[Dict[str, Nullable[Any]]]
inflection: NotRequired[Dict[str, Nullable[Any]]]
inocloud: NotRequired[Dict[str, Nullable[Any]]]
io_net: NotRequired[Dict[str, Nullable[Any]]]
ionstream: NotRequired[Dict[str, Nullable[Any]]]
klusterai: NotRequired[Dict[str, Nullable[Any]]]
lambda_: NotRequired[Dict[str, Nullable[Any]]]
lepton: NotRequired[Dict[str, Nullable[Any]]]
liquid: NotRequired[Dict[str, Nullable[Any]]]
lynn: NotRequired[Dict[str, Nullable[Any]]]
lynn_private: NotRequired[Dict[str, Nullable[Any]]]
mancer: NotRequired[Dict[str, Nullable[Any]]]
mancer_old: NotRequired[Dict[str, Nullable[Any]]]
mara: NotRequired[Dict[str, Nullable[Any]]]
meta: NotRequired[Dict[str, Nullable[Any]]]
minimax: NotRequired[Dict[str, Nullable[Any]]]
mistral: NotRequired[Dict[str, Nullable[Any]]]
modal: NotRequired[Dict[str, Nullable[Any]]]
modelrun: NotRequired[Dict[str, Nullable[Any]]]
modular: NotRequired[Dict[str, Nullable[Any]]]
moonshotai: NotRequired[Dict[str, Nullable[Any]]]
morph: NotRequired[Dict[str, Nullable[Any]]]
ncompass: NotRequired[Dict[str, Nullable[Any]]]
nebius: NotRequired[Dict[str, Nullable[Any]]]
nex_agi: NotRequired[Dict[str, Nullable[Any]]]
nextbit: NotRequired[Dict[str, Nullable[Any]]]
nineteen: NotRequired[Dict[str, Nullable[Any]]]
novita: NotRequired[Dict[str, Nullable[Any]]]
nvidia: NotRequired[Dict[str, Nullable[Any]]]
octoai: NotRequired[Dict[str, Nullable[Any]]]
open_inference: NotRequired[Dict[str, Nullable[Any]]]
openai: NotRequired[Dict[str, Nullable[Any]]]
parasail: NotRequired[Dict[str, Nullable[Any]]]
perceptron: NotRequired[Dict[str, Nullable[Any]]]
perplexity: NotRequired[Dict[str, Nullable[Any]]]
phala: NotRequired[Dict[str, Nullable[Any]]]
poolside: NotRequired[Dict[str, Nullable[Any]]]
quiver: NotRequired[Dict[str, Nullable[Any]]]
recraft: NotRequired[Dict[str, Nullable[Any]]]
recursal: NotRequired[Dict[str, Nullable[Any]]]
reflection: NotRequired[Dict[str, Nullable[Any]]]
reka: NotRequired[Dict[str, Nullable[Any]]]
relace: NotRequired[Dict[str, Nullable[Any]]]
replicate: NotRequired[Dict[str, Nullable[Any]]]
sakana_ai: NotRequired[Dict[str, Nullable[Any]]]
sambanova: NotRequired[Dict[str, Nullable[Any]]]
sambanova_cloaked: NotRequired[Dict[str, Nullable[Any]]]
seed: NotRequired[Dict[str, Nullable[Any]]]
sf_compute: NotRequired[Dict[str, Nullable[Any]]]
siliconflow: NotRequired[Dict[str, Nullable[Any]]]
sourceful: NotRequired[Dict[str, Nullable[Any]]]
stealth: NotRequired[Dict[str, Nullable[Any]]]
stepfun: NotRequired[Dict[str, Nullable[Any]]]
streamlake: NotRequired[Dict[str, Nullable[Any]]]
switchpoint: NotRequired[Dict[str, Nullable[Any]]]
targon: NotRequired[Dict[str, Nullable[Any]]]
tenstorrent: NotRequired[Dict[str, Nullable[Any]]]
together: NotRequired[Dict[str, Nullable[Any]]]
together_lite: NotRequired[Dict[str, Nullable[Any]]]
ubicloud: NotRequired[Dict[str, Nullable[Any]]]
upstage: NotRequired[Dict[str, Nullable[Any]]]
venice: NotRequired[Dict[str, Nullable[Any]]]
wafer: NotRequired[Dict[str, Nullable[Any]]]
wandb: NotRequired[Dict[str, Nullable[Any]]]
xai: NotRequired[Dict[str, Nullable[Any]]]
xiaomi: NotRequired[Dict[str, Nullable[Any]]]
z_ai: NotRequired[Dict[str, Nullable[Any]]]
class ImageGenerationRequestOptions(BaseModel):
r"""Provider-specific options keyed by provider slug. Only options for the matched provider are forwarded; the rest are ignored. Unrecognized keys are silently dropped."""
oneai: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="01ai")
] = None
ai21: Optional[Dict[str, Nullable[Any]]] = None
aion_labs: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="aion-labs")
] = None
akashml: Optional[Dict[str, Nullable[Any]]] = None
alibaba: Optional[Dict[str, Nullable[Any]]] = None
amazon_bedrock: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="amazon-bedrock")
] = None
amazon_nova: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="amazon-nova")
] = None
ambient: Optional[Dict[str, Nullable[Any]]] = None
anthropic: Optional[Dict[str, Nullable[Any]]] = None
anyscale: Optional[Dict[str, Nullable[Any]]] = None
arcee_ai: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="arcee-ai")
] = None
atlas_cloud: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="atlas-cloud")
] = None
atoma: Optional[Dict[str, Nullable[Any]]] = None
avian: Optional[Dict[str, Nullable[Any]]] = None
azure: Optional[Dict[str, Nullable[Any]]] = None
baidu: Optional[Dict[str, Nullable[Any]]] = None
baseten: Optional[Dict[str, Nullable[Any]]] = None
black_forest_labs: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="black-forest-labs")
] = None
byteplus: Optional[Dict[str, Nullable[Any]]] = None
centml: Optional[Dict[str, Nullable[Any]]] = None
cerebras: Optional[Dict[str, Nullable[Any]]] = None
chutes: Optional[Dict[str, Nullable[Any]]] = None
cirrascale: Optional[Dict[str, Nullable[Any]]] = None
clarifai: Optional[Dict[str, Nullable[Any]]] = None
cloudflare: Optional[Dict[str, Nullable[Any]]] = None
cohere: Optional[Dict[str, Nullable[Any]]] = None
crofai: Optional[Dict[str, Nullable[Any]]] = None
crucible: Optional[Dict[str, Nullable[Any]]] = None
crusoe: Optional[Dict[str, Nullable[Any]]] = None
darkbloom: Optional[Dict[str, Nullable[Any]]] = None
decart: Optional[Dict[str, Nullable[Any]]] = None
deepinfra: Optional[Dict[str, Nullable[Any]]] = None
deepseek: Optional[Dict[str, Nullable[Any]]] = None
dekallm: Optional[Dict[str, Nullable[Any]]] = None
digitalocean: Optional[Dict[str, Nullable[Any]]] = None
enfer: Optional[Dict[str, Nullable[Any]]] = None
fake_provider: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="fake-provider")
] = None
featherless: Optional[Dict[str, Nullable[Any]]] = None
fireworks: Optional[Dict[str, Nullable[Any]]] = None
friendli: Optional[Dict[str, Nullable[Any]]] = None
gmicloud: Optional[Dict[str, Nullable[Any]]] = None
google_ai_studio: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="google-ai-studio")
] = None
google_vertex: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="google-vertex")
] = None
gopomelo: Optional[Dict[str, Nullable[Any]]] = None
groq: Optional[Dict[str, Nullable[Any]]] = None
heygen: Optional[Dict[str, Nullable[Any]]] = None
huggingface: Optional[Dict[str, Nullable[Any]]] = None
hyperbolic: Optional[Dict[str, Nullable[Any]]] = None
hyperbolic_quantized: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="hyperbolic-quantized")
] = None
inception: Optional[Dict[str, Nullable[Any]]] = None
inceptron: Optional[Dict[str, Nullable[Any]]] = None
inferact_vllm: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="inferact-vllm")
] = None
inference_net: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="inference-net")
] = None
infermatic: Optional[Dict[str, Nullable[Any]]] = None
inflection: Optional[Dict[str, Nullable[Any]]] = None
inocloud: Optional[Dict[str, Nullable[Any]]] = None
io_net: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="io-net")
] = None
ionstream: Optional[Dict[str, Nullable[Any]]] = None
klusterai: Optional[Dict[str, Nullable[Any]]] = None
lambda_: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="lambda")
] = None
lepton: Optional[Dict[str, Nullable[Any]]] = None
liquid: Optional[Dict[str, Nullable[Any]]] = None
lynn: Optional[Dict[str, Nullable[Any]]] = None
lynn_private: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="lynn-private")
] = None
mancer: Optional[Dict[str, Nullable[Any]]] = None
mancer_old: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="mancer-old")
] = None
mara: Optional[Dict[str, Nullable[Any]]] = None
meta: Optional[Dict[str, Nullable[Any]]] = None
minimax: Optional[Dict[str, Nullable[Any]]] = None
mistral: Optional[Dict[str, Nullable[Any]]] = None
modal: Optional[Dict[str, Nullable[Any]]] = None
modelrun: Optional[Dict[str, Nullable[Any]]] = None
modular: Optional[Dict[str, Nullable[Any]]] = None
moonshotai: Optional[Dict[str, Nullable[Any]]] = None
morph: Optional[Dict[str, Nullable[Any]]] = None
ncompass: Optional[Dict[str, Nullable[Any]]] = None
nebius: Optional[Dict[str, Nullable[Any]]] = None
nex_agi: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="nex-agi")
] = None
nextbit: Optional[Dict[str, Nullable[Any]]] = None
nineteen: Optional[Dict[str, Nullable[Any]]] = None
novita: Optional[Dict[str, Nullable[Any]]] = None
nvidia: Optional[Dict[str, Nullable[Any]]] = None
octoai: Optional[Dict[str, Nullable[Any]]] = None
open_inference: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="open-inference")
] = None
openai: Optional[Dict[str, Nullable[Any]]] = None
parasail: Optional[Dict[str, Nullable[Any]]] = None
perceptron: Optional[Dict[str, Nullable[Any]]] = None
perplexity: Optional[Dict[str, Nullable[Any]]] = None
phala: Optional[Dict[str, Nullable[Any]]] = None
poolside: Optional[Dict[str, Nullable[Any]]] = None
quiver: Optional[Dict[str, Nullable[Any]]] = None
recraft: Optional[Dict[str, Nullable[Any]]] = None
recursal: Optional[Dict[str, Nullable[Any]]] = None
reflection: Optional[Dict[str, Nullable[Any]]] = None
reka: Optional[Dict[str, Nullable[Any]]] = None
relace: Optional[Dict[str, Nullable[Any]]] = None
replicate: Optional[Dict[str, Nullable[Any]]] = None
sakana_ai: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="sakana-ai")
] = None
sambanova: Optional[Dict[str, Nullable[Any]]] = None
sambanova_cloaked: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="sambanova-cloaked")
] = None
seed: Optional[Dict[str, Nullable[Any]]] = None
sf_compute: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="sf-compute")
] = None
siliconflow: Optional[Dict[str, Nullable[Any]]] = None
sourceful: Optional[Dict[str, Nullable[Any]]] = None
stealth: Optional[Dict[str, Nullable[Any]]] = None
stepfun: Optional[Dict[str, Nullable[Any]]] = None
streamlake: Optional[Dict[str, Nullable[Any]]] = None
switchpoint: Optional[Dict[str, Nullable[Any]]] = None
targon: Optional[Dict[str, Nullable[Any]]] = None
tenstorrent: Optional[Dict[str, Nullable[Any]]] = None
together: Optional[Dict[str, Nullable[Any]]] = None
together_lite: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="together-lite")
] = None
ubicloud: Optional[Dict[str, Nullable[Any]]] = None
upstage: Optional[Dict[str, Nullable[Any]]] = None
venice: Optional[Dict[str, Nullable[Any]]] = None
wafer: Optional[Dict[str, Nullable[Any]]] = None
wandb: Optional[Dict[str, Nullable[Any]]] = None
xai: Optional[Dict[str, Nullable[Any]]] = None
xiaomi: Optional[Dict[str, Nullable[Any]]] = None
z_ai: Annotated[
Optional[Dict[str, Nullable[Any]]], pydantic.Field(alias="z-ai")
] = None
class ImageGenerationRequestProviderTypedDict(TypedDict):
r"""Provider-specific passthrough configuration"""
options: NotRequired[ImageGenerationRequestOptionsTypedDict]
class ImageGenerationRequestProvider(BaseModel):
r"""Provider-specific passthrough configuration"""
options: Optional[ImageGenerationRequestOptions] = None
ImageGenerationRequestQuality = Union[
Literal[
"auto",
"low",
"medium",
"high",
],
UnrecognizedStr,
]
r"""Rendering quality. Providers without a quality knob ignore this."""
ImageGenerationRequestResolution = Union[
Literal[
"512",
"1K",
"2K",
"4K",
],
UnrecognizedStr,
]
r"""Normalized resolution tier of the generated image. Concrete pixel dimensions are derived per-provider."""
class ImageGenerationRequestTypedDict(TypedDict):
r"""Image generation request input"""
model: str
r"""The image generation model to use"""
prompt: str
r"""Text description of the desired image"""
aspect_ratio: NotRequired[ImageGenerationRequestAspectRatio]
r"""Normalized aspect ratio of the generated image. Providers clamp to their supported subset."""
background: NotRequired[ImageGenerationRequestBackground]
r"""Background treatment. `transparent` requires an output_format that supports alpha (png or webp)."""
input_references: NotRequired[List[ContentPartImageTypedDict]]
r"""Reference images to guide image-to-image generation, as base64 data URLs or HTTP(S) URLs."""
n: NotRequired[int]
r"""Number of images to generate (1-10). Providers that only support single-image generation reject n > 1."""
output_compression: NotRequired[int]
r"""Compression level (0-100) for webp/jpeg output. Ignored for png and by providers without a compression knob."""
output_format: NotRequired[ImageGenerationRequestOutputFormat]
r"""Encoding of the returned image bytes. Most models produce raster formats (png, jpeg, webp). SVG is supported by vectorization models (e.g. Quiver) — the SVG markup is UTF-8 base64-encoded in `b64_json`."""
provider: NotRequired[ImageGenerationRequestProviderTypedDict]
r"""Provider-specific passthrough configuration"""
quality: NotRequired[ImageGenerationRequestQuality]
r"""Rendering quality. Providers without a quality knob ignore this."""
resolution: NotRequired[ImageGenerationRequestResolution]
r"""Normalized resolution tier of the generated image. Concrete pixel dimensions are derived per-provider."""
seed: NotRequired[int]
r"""If specified, the generation will sample deterministically, such that repeated requests with the same seed and parameters should return the same result. Determinism is not guaranteed for all providers."""
size: NotRequired[str]
r"""Optional. A convenience shorthand for output dimensions — pass a tier (\"2K\", \"4K\") or explicit pixels (\"2048x2048\") and we normalize it to the right dimensions for the chosen provider. Interchangeable with resolution + aspect_ratio; use those directly for enumerated, per-model discoverable values. Conflicting size + resolution/aspect_ratio is rejected."""
stream: NotRequired[bool]
r"""If true, partial images are streamed as SSE events as they become available. Only supported by providers with native streaming (currently OpenAI). Non-streaming providers ignore this flag and return a buffered response."""
class ImageGenerationRequest(BaseModel):
r"""Image generation request input"""
model: str
r"""The image generation model to use"""
prompt: str
r"""Text description of the desired image"""
aspect_ratio: Annotated[
Optional[ImageGenerationRequestAspectRatio],
PlainValidator(validate_open_enum(False)),
] = None
r"""Normalized aspect ratio of the generated image. Providers clamp to their supported subset."""
background: Annotated[
Optional[ImageGenerationRequestBackground],
PlainValidator(validate_open_enum(False)),
] = None
r"""Background treatment. `transparent` requires an output_format that supports alpha (png or webp)."""
input_references: Optional[List[ContentPartImage]] = None
r"""Reference images to guide image-to-image generation, as base64 data URLs or HTTP(S) URLs."""
n: Optional[int] = None
r"""Number of images to generate (1-10). Providers that only support single-image generation reject n > 1."""
output_compression: Optional[int] = None
r"""Compression level (0-100) for webp/jpeg output. Ignored for png and by providers without a compression knob."""
output_format: Annotated[
Optional[ImageGenerationRequestOutputFormat],
PlainValidator(validate_open_enum(False)),
] = None
r"""Encoding of the returned image bytes. Most models produce raster formats (png, jpeg, webp). SVG is supported by vectorization models (e.g. Quiver) — the SVG markup is UTF-8 base64-encoded in `b64_json`."""
provider: Optional[ImageGenerationRequestProvider] = None
r"""Provider-specific passthrough configuration"""
quality: Annotated[
Optional[ImageGenerationRequestQuality],
PlainValidator(validate_open_enum(False)),
] = None
r"""Rendering quality. Providers without a quality knob ignore this."""
resolution: Annotated[
Optional[ImageGenerationRequestResolution],
PlainValidator(validate_open_enum(False)),
] = None
r"""Normalized resolution tier of the generated image. Concrete pixel dimensions are derived per-provider."""
seed: Optional[int] = None
r"""If specified, the generation will sample deterministically, such that repeated requests with the same seed and parameters should return the same result. Determinism is not guaranteed for all providers."""
size: Optional[str] = None
r"""Optional. A convenience shorthand for output dimensions — pass a tier (\"2K\", \"4K\") or explicit pixels (\"2048x2048\") and we normalize it to the right dimensions for the chosen provider. Interchangeable with resolution + aspect_ratio; use those directly for enumerated, per-model discoverable values. Conflicting size + resolution/aspect_ratio is rejected."""
stream: Optional[bool] = None
r"""If true, partial images are streamed as SSE events as they become available. Only supported by providers with native streaming (currently OpenAI). Non-streaming providers ignore this flag and return a buffered response."""
@@ -0,0 +1,46 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .imagegenerationusage import ImageGenerationUsage, ImageGenerationUsageTypedDict
from openrouter.types import BaseModel
from typing import List, Optional
from typing_extensions import NotRequired, TypedDict
class ImageGenerationResponseDataTypedDict(TypedDict):
b64_json: str
r"""Base64-encoded image bytes"""
media_type: NotRequired[str]
r"""Media type (MIME type) of the image. Omitted when the output is a standard raster format (PNG). Present for non-raster outputs such as SVG (`image/svg+xml`)."""
class ImageGenerationResponseData(BaseModel):
b64_json: str
r"""Base64-encoded image bytes"""
media_type: Optional[str] = None
r"""Media type (MIME type) of the image. Omitted when the output is a standard raster format (PNG). Present for non-raster outputs such as SVG (`image/svg+xml`)."""
class ImageGenerationResponseTypedDict(TypedDict):
r"""Image generation response"""
created: int
r"""Unix timestamp (seconds) when the image was generated"""
data: List[ImageGenerationResponseDataTypedDict]
r"""Generated images"""
usage: NotRequired[ImageGenerationUsageTypedDict]
r"""Token and cost usage for the image generation request, when available"""
class ImageGenerationResponse(BaseModel):
r"""Image generation response"""
created: int
r"""Unix timestamp (seconds) when the image was generated"""
data: List[ImageGenerationResponseData]
r"""Generated images"""
usage: Optional[ImageGenerationUsage] = None
r"""Token and cost usage for the image generation request, when available"""
@@ -16,7 +16,7 @@ from typing import Literal, Optional, Union
from typing_extensions import Annotated, NotRequired, TypedDict
Background = Union[
ImageGenerationServerToolBackground = Union[
Literal[
"transparent",
"opaque",
@@ -64,7 +64,7 @@ Moderation = Union[
]
OutputFormat = Union[
ImageGenerationServerToolOutputFormat = Union[
Literal[
"png",
"webp",
@@ -74,7 +74,7 @@ OutputFormat = Union[
]
Quality = Union[
ImageGenerationServerToolQuality = Union[
Literal[
"low",
"medium",
@@ -103,15 +103,15 @@ class ImageGenerationServerToolTypedDict(TypedDict):
r"""Image generation tool configuration"""
type: ImageGenerationServerToolType
background: NotRequired[Background]
background: NotRequired[ImageGenerationServerToolBackground]
input_fidelity: NotRequired[Nullable[InputFidelity]]
input_image_mask: NotRequired[InputImageMaskTypedDict]
model: NotRequired[ModelEnum]
moderation: NotRequired[Moderation]
output_compression: NotRequired[int]
output_format: NotRequired[OutputFormat]
output_format: NotRequired[ImageGenerationServerToolOutputFormat]
partial_images: NotRequired[int]
quality: NotRequired[Quality]
quality: NotRequired[ImageGenerationServerToolQuality]
size: NotRequired[Size]
@@ -121,7 +121,8 @@ class ImageGenerationServerTool(BaseModel):
type: ImageGenerationServerToolType
background: Annotated[
Optional[Background], PlainValidator(validate_open_enum(False))
Optional[ImageGenerationServerToolBackground],
PlainValidator(validate_open_enum(False)),
] = None
input_fidelity: Annotated[
@@ -141,14 +142,16 @@ class ImageGenerationServerTool(BaseModel):
output_compression: Optional[int] = None
output_format: Annotated[
Optional[OutputFormat], PlainValidator(validate_open_enum(False))
Optional[ImageGenerationServerToolOutputFormat],
PlainValidator(validate_open_enum(False)),
] = None
partial_images: Optional[int] = None
quality: Annotated[Optional[Quality], PlainValidator(validate_open_enum(False))] = (
None
)
quality: Annotated[
Optional[ImageGenerationServerToolQuality],
PlainValidator(validate_open_enum(False)),
] = None
size: Annotated[Optional[Size], PlainValidator(validate_open_enum(False))] = None
@@ -0,0 +1,331 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .anthropicspeed import AnthropicSpeed
from .anthropicusageiteration import (
AnthropicUsageIteration,
AnthropicUsageIterationTypedDict,
)
from .costdetails import CostDetails, CostDetailsTypedDict
from openrouter.types import (
BaseModel,
Nullable,
OptionalNullable,
UNSET,
UNSET_SENTINEL,
)
from openrouter.utils import validate_open_enum
from pydantic import model_serializer
from pydantic.functional_validators import PlainValidator
from typing import List, Optional
from typing_extensions import Annotated, NotRequired, TypedDict
class ImageGenerationUsageCompletionTokensDetailsTypedDict(TypedDict):
audio_tokens: NotRequired[Nullable[int]]
r"""Tokens generated by the model for audio output."""
image_tokens: NotRequired[Nullable[int]]
r"""Tokens generated by the model for image output."""
reasoning_tokens: NotRequired[Nullable[int]]
r"""Tokens generated by the model for reasoning."""
class ImageGenerationUsageCompletionTokensDetails(BaseModel):
audio_tokens: OptionalNullable[int] = UNSET
r"""Tokens generated by the model for audio output."""
image_tokens: OptionalNullable[int] = UNSET
r"""Tokens generated by the model for image output."""
reasoning_tokens: OptionalNullable[int] = UNSET
r"""Tokens generated by the model for reasoning."""
@model_serializer(mode="wrap")
def serialize_model(self, handler):
optional_fields = ["audio_tokens", "image_tokens", "reasoning_tokens"]
nullable_fields = ["audio_tokens", "image_tokens", "reasoning_tokens"]
null_default_fields = []
serialized = handler(self)
m = {}
for n, f in type(self).model_fields.items():
k = f.alias or n
val = serialized.get(k)
serialized.pop(k, None)
optional_nullable = k in optional_fields and k in nullable_fields
is_set = (
self.__pydantic_fields_set__.intersection({n})
or k in null_default_fields
) # pylint: disable=no-member
if val is not None and val != UNSET_SENTINEL:
m[k] = val
elif val != UNSET_SENTINEL and (
not k in optional_fields or (optional_nullable and is_set)
):
m[k] = val
return m
class ImageGenerationUsagePromptTokensDetailsTypedDict(TypedDict):
r"""Breakdown of tokens used in the prompt."""
audio_tokens: NotRequired[Nullable[int]]
r"""Tokens used for input audio."""
cache_write_tokens: NotRequired[Nullable[int]]
r"""Tokens written to cache. Only returned for models with explicit caching and cache write pricing."""
cached_tokens: NotRequired[Nullable[int]]
r"""Tokens cached by the endpoint."""
file_tokens: NotRequired[Nullable[int]]
r"""Tokens used for input files/documents."""
video_tokens: NotRequired[Nullable[int]]
r"""Tokens used for input video."""
class ImageGenerationUsagePromptTokensDetails(BaseModel):
r"""Breakdown of tokens used in the prompt."""
audio_tokens: OptionalNullable[int] = UNSET
r"""Tokens used for input audio."""
cache_write_tokens: OptionalNullable[int] = UNSET
r"""Tokens written to cache. Only returned for models with explicit caching and cache write pricing."""
cached_tokens: OptionalNullable[int] = UNSET
r"""Tokens cached by the endpoint."""
file_tokens: OptionalNullable[int] = UNSET
r"""Tokens used for input files/documents."""
video_tokens: OptionalNullable[int] = UNSET
r"""Tokens used for input video."""
@model_serializer(mode="wrap")
def serialize_model(self, handler):
optional_fields = [
"audio_tokens",
"cache_write_tokens",
"cached_tokens",
"file_tokens",
"video_tokens",
]
nullable_fields = [
"audio_tokens",
"cache_write_tokens",
"cached_tokens",
"file_tokens",
"video_tokens",
]
null_default_fields = []
serialized = handler(self)
m = {}
for n, f in type(self).model_fields.items():
k = f.alias or n
val = serialized.get(k)
serialized.pop(k, None)
optional_nullable = k in optional_fields and k in nullable_fields
is_set = (
self.__pydantic_fields_set__.intersection({n})
or k in null_default_fields
) # pylint: disable=no-member
if val is not None and val != UNSET_SENTINEL:
m[k] = val
elif val != UNSET_SENTINEL and (
not k in optional_fields or (optional_nullable and is_set)
):
m[k] = val
return m
class ServerToolUseTypedDict(TypedDict):
r"""Usage for server-side tool execution (e.g., web search)"""
tool_calls_executed: NotRequired[Nullable[int]]
r"""Number of OpenRouter server tool calls that executed and produced a result."""
tool_calls_requested: NotRequired[Nullable[int]]
r"""Total number of OpenRouter server-orchestrated tool calls the model requested, across all tool types. Provider-native tools (e.g. native web search) are not counted here."""
web_search_requests: NotRequired[Nullable[int]]
r"""Number of web searches performed by server-side tools. For server-orchestrated tool calls a web search is also counted in tool_calls_requested; provider-native web search may report web_search_requests only. Do not sum the two."""
class ServerToolUse(BaseModel):
r"""Usage for server-side tool execution (e.g., web search)"""
tool_calls_executed: OptionalNullable[int] = UNSET
r"""Number of OpenRouter server tool calls that executed and produced a result."""
tool_calls_requested: OptionalNullable[int] = UNSET
r"""Total number of OpenRouter server-orchestrated tool calls the model requested, across all tool types. Provider-native tools (e.g. native web search) are not counted here."""
web_search_requests: OptionalNullable[int] = UNSET
r"""Number of web searches performed by server-side tools. For server-orchestrated tool calls a web search is also counted in tool_calls_requested; provider-native web search may report web_search_requests only. Do not sum the two."""
@model_serializer(mode="wrap")
def serialize_model(self, handler):
optional_fields = [
"tool_calls_executed",
"tool_calls_requested",
"web_search_requests",
]
nullable_fields = [
"tool_calls_executed",
"tool_calls_requested",
"web_search_requests",
]
null_default_fields = []
serialized = handler(self)
m = {}
for n, f in type(self).model_fields.items():
k = f.alias or n
val = serialized.get(k)
serialized.pop(k, None)
optional_nullable = k in optional_fields and k in nullable_fields
is_set = (
self.__pydantic_fields_set__.intersection({n})
or k in null_default_fields
) # pylint: disable=no-member
if val is not None and val != UNSET_SENTINEL:
m[k] = val
elif val != UNSET_SENTINEL and (
not k in optional_fields or (optional_nullable and is_set)
):
m[k] = val
return m
class ImageGenerationUsageTypedDict(TypedDict):
r"""Token and cost usage for the image generation request, when available"""
completion_tokens: int
r"""The tokens generated"""
prompt_tokens: int
r"""Including images, input audio, and tools if any"""
total_tokens: int
r"""Sum of the above two fields"""
completion_tokens_details: NotRequired[
Nullable[ImageGenerationUsageCompletionTokensDetailsTypedDict]
]
cost: NotRequired[Nullable[float]]
r"""Cost of the completion"""
cost_details: NotRequired[Nullable[CostDetailsTypedDict]]
r"""Breakdown of upstream inference costs"""
is_byok: NotRequired[bool]
r"""Whether a request was made using a Bring Your Own Key configuration"""
iterations: NotRequired[Nullable[List[AnthropicUsageIterationTypedDict]]]
prompt_tokens_details: NotRequired[
Nullable[ImageGenerationUsagePromptTokensDetailsTypedDict]
]
r"""Breakdown of tokens used in the prompt."""
server_tool_use: NotRequired[Nullable[ServerToolUseTypedDict]]
r"""Usage for server-side tool execution (e.g., web search)"""
service_tier: NotRequired[Nullable[str]]
r"""The service tier used by the upstream provider for this request"""
speed: NotRequired[Nullable[AnthropicSpeed]]
class ImageGenerationUsage(BaseModel):
r"""Token and cost usage for the image generation request, when available"""
completion_tokens: int
r"""The tokens generated"""
prompt_tokens: int
r"""Including images, input audio, and tools if any"""
total_tokens: int
r"""Sum of the above two fields"""
completion_tokens_details: OptionalNullable[
ImageGenerationUsageCompletionTokensDetails
] = UNSET
cost: OptionalNullable[float] = UNSET
r"""Cost of the completion"""
cost_details: OptionalNullable[CostDetails] = UNSET
r"""Breakdown of upstream inference costs"""
is_byok: Optional[bool] = None
r"""Whether a request was made using a Bring Your Own Key configuration"""
iterations: OptionalNullable[List[AnthropicUsageIteration]] = UNSET
prompt_tokens_details: OptionalNullable[ImageGenerationUsagePromptTokensDetails] = (
UNSET
)
r"""Breakdown of tokens used in the prompt."""
server_tool_use: OptionalNullable[ServerToolUse] = UNSET
r"""Usage for server-side tool execution (e.g., web search)"""
service_tier: OptionalNullable[str] = UNSET
r"""The service tier used by the upstream provider for this request"""
speed: Annotated[
OptionalNullable[AnthropicSpeed], PlainValidator(validate_open_enum(False))
] = UNSET
@model_serializer(mode="wrap")
def serialize_model(self, handler):
optional_fields = [
"completion_tokens_details",
"cost",
"cost_details",
"is_byok",
"iterations",
"prompt_tokens_details",
"server_tool_use",
"service_tier",
"speed",
]
nullable_fields = [
"completion_tokens_details",
"cost",
"cost_details",
"iterations",
"prompt_tokens_details",
"server_tool_use",
"service_tier",
"speed",
]
null_default_fields = []
serialized = handler(self)
m = {}
for n, f in type(self).model_fields.items():
k = f.alias or n
val = serialized.get(k)
serialized.pop(k, None)
optional_nullable = k in optional_fields and k in nullable_fields
is_set = (
self.__pydantic_fields_set__.intersection({n})
or k in null_default_fields
) # pylint: disable=no-member
if val is not None and val != UNSET_SENTINEL:
m[k] = val
elif val != UNSET_SENTINEL and (
not k in optional_fields or (optional_nullable and is_set)
):
m[k] = val
return m
@@ -0,0 +1,34 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import BaseModel
from typing import Literal
from typing_extensions import TypedDict
ImageGenPartialImageEventType = Literal["image_generation.partial_image",]
r"""The event type"""
class ImageGenPartialImageEventTypedDict(TypedDict):
r"""Emitted when a partial image becomes available during streaming generation"""
b64_json: str
r"""Base64-encoded partial image data"""
partial_image_index: int
r"""0-based index indicating which partial image this is in the sequence"""
type: ImageGenPartialImageEventType
r"""The event type"""
class ImageGenPartialImageEvent(BaseModel):
r"""Emitted when a partial image becomes available during streaming generation"""
b64_json: str
r"""Base64-encoded partial image data"""
partial_image_index: int
r"""0-based index indicating which partial image this is in the sequence"""
type: ImageGenPartialImageEventType
r"""The event type"""
@@ -0,0 +1,95 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import (
BaseModel,
Nullable,
OptionalNullable,
UNSET,
UNSET_SENTINEL,
)
from pydantic import model_serializer
from typing import Literal
from typing_extensions import NotRequired, TypedDict
class ImageGenStreamErrorEventErrorTypedDict(TypedDict):
r"""Provider error details"""
message: str
r"""Provider error message"""
code: NotRequired[Nullable[str]]
r"""Provider error code, when supplied"""
param: NotRequired[Nullable[str]]
r"""Request parameter associated with the error, when supplied"""
type: NotRequired[Nullable[str]]
r"""Provider error type, when supplied"""
class ImageGenStreamErrorEventError(BaseModel):
r"""Provider error details"""
message: str
r"""Provider error message"""
code: OptionalNullable[str] = UNSET
r"""Provider error code, when supplied"""
param: OptionalNullable[str] = UNSET
r"""Request parameter associated with the error, when supplied"""
type: OptionalNullable[str] = UNSET
r"""Provider error type, when supplied"""
@model_serializer(mode="wrap")
def serialize_model(self, handler):
optional_fields = ["code", "param", "type"]
nullable_fields = ["code", "param", "type"]
null_default_fields = []
serialized = handler(self)
m = {}
for n, f in type(self).model_fields.items():
k = f.alias or n
val = serialized.get(k)
serialized.pop(k, None)
optional_nullable = k in optional_fields and k in nullable_fields
is_set = (
self.__pydantic_fields_set__.intersection({n})
or k in null_default_fields
) # pylint: disable=no-member
if val is not None and val != UNSET_SENTINEL:
m[k] = val
elif val != UNSET_SENTINEL and (
not k in optional_fields or (optional_nullable and is_set)
):
m[k] = val
return m
ImageGenStreamErrorEventType = Literal["error",]
r"""The event type"""
class ImageGenStreamErrorEventTypedDict(TypedDict):
r"""Emitted when streaming generation fails after the SSE response starts"""
error: ImageGenStreamErrorEventErrorTypedDict
r"""Provider error details"""
type: ImageGenStreamErrorEventType
r"""The event type"""
class ImageGenStreamErrorEvent(BaseModel):
r"""Emitted when streaming generation fails after the SSE response starts"""
error: ImageGenStreamErrorEventError
r"""Provider error details"""
type: ImageGenStreamErrorEventType
r"""The event type"""
@@ -0,0 +1,29 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .imageoutputmodality import ImageOutputModality
from .inputmodality import InputModality
from openrouter.types import BaseModel
from openrouter.utils import validate_open_enum
from pydantic.functional_validators import PlainValidator
from typing import List
from typing_extensions import Annotated, TypedDict
class ImageModelArchitectureTypedDict(TypedDict):
input_modalities: List[InputModality]
r"""Supported input modalities"""
output_modalities: List[ImageOutputModality]
r"""Supported output modalities"""
class ImageModelArchitecture(BaseModel):
input_modalities: List[
Annotated[InputModality, PlainValidator(validate_open_enum(False))]
]
r"""Supported input modalities"""
output_modalities: List[
Annotated[ImageOutputModality, PlainValidator(validate_open_enum(False))]
]
r"""Supported output modalities"""
@@ -0,0 +1,24 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .imageendpoint import ImageEndpoint, ImageEndpointTypedDict
from openrouter.types import BaseModel
from typing import List
from typing_extensions import TypedDict
class ImageModelEndpointsResponseTypedDict(TypedDict):
r"""The full per-endpoint records for an image model."""
endpoints: List[ImageEndpointTypedDict]
id: str
r"""Model slug"""
class ImageModelEndpointsResponse(BaseModel):
r"""The full per-endpoint records for an image model."""
endpoints: List[ImageEndpoint]
id: str
r"""Model slug"""
@@ -0,0 +1,56 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .capabilitydescriptor import CapabilityDescriptor, CapabilityDescriptorTypedDict
from .imagemodelarchitecture import (
ImageModelArchitecture,
ImageModelArchitectureTypedDict,
)
from openrouter.types import BaseModel
from typing import Dict
from typing_extensions import TypedDict
class ImageModelListItemTypedDict(TypedDict):
r"""A single image model in the discovery listing."""
architecture: ImageModelArchitectureTypedDict
created: int
r"""Unix timestamp (seconds) of when the model was created"""
description: str
endpoints: str
r"""Relative URL to the full per-endpoint records for this model"""
id: str
r"""Model slug"""
name: str
r"""Display name"""
supported_parameters: Dict[str, CapabilityDescriptorTypedDict]
r"""Union of supported parameters across every endpoint of this model. Coarse discovery aid; the definitive per-endpoint set is behind the endpoints URL."""
supports_streaming: bool
r"""Whether any endpoint of this model supports native SSE streaming on the dedicated Image API (i.e. `stream: true` in the request). OR across endpoints."""
class ImageModelListItem(BaseModel):
r"""A single image model in the discovery listing."""
architecture: ImageModelArchitecture
created: int
r"""Unix timestamp (seconds) of when the model was created"""
description: str
endpoints: str
r"""Relative URL to the full per-endpoint records for this model"""
id: str
r"""Model slug"""
name: str
r"""Display name"""
supported_parameters: Dict[str, CapabilityDescriptor]
r"""Union of supported parameters across every endpoint of this model. Coarse discovery aid; the definitive per-endpoint set is behind the endpoints URL."""
supports_streaming: bool
r"""Whether any endpoint of this model supports native SSE streaming on the dedicated Image API (i.e. `stream: true` in the request). OR across endpoints."""
@@ -0,0 +1,19 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .imagemodellistitem import ImageModelListItem, ImageModelListItemTypedDict
from openrouter.types import BaseModel
from typing import List
from typing_extensions import TypedDict
class ImageModelsListResponseTypedDict(TypedDict):
r"""List of image generation models."""
data: List[ImageModelListItemTypedDict]
class ImageModelsListResponse(BaseModel):
r"""List of image generation models."""
data: List[ImageModelListItem]
@@ -0,0 +1,20 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import UnrecognizedStr
from typing import Literal, Union
ImageOutputModality = Union[
Literal[
"text",
"image",
"embeddings",
"audio",
"video",
"rerank",
"speech",
"transcription",
],
UnrecognizedStr,
]
@@ -0,0 +1,51 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import BaseModel, UnrecognizedStr
from openrouter.utils import validate_open_enum
from pydantic.functional_validators import PlainValidator
from typing import Literal, Optional, Union
from typing_extensions import Annotated, NotRequired, TypedDict
Billable = Union[
Literal[
"output_image",
"input_image",
"input_font",
"input_reference",
"input_text",
],
UnrecognizedStr,
]
Unit = Union[
Literal[
"image",
"megapixel",
"token",
],
UnrecognizedStr,
]
class ImagePricingEntryTypedDict(TypedDict):
r"""One billable pricing line for an image provider."""
billable: Billable
cost_usd: float
unit: Unit
variant: NotRequired[str]
class ImagePricingEntry(BaseModel):
r"""One billable pricing line for an image provider."""
billable: Annotated[Billable, PlainValidator(validate_open_enum(False))]
cost_usd: float
unit: Annotated[Unit, PlainValidator(validate_open_enum(False))]
variant: Optional[str] = None
@@ -0,0 +1,48 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .imagegencompletedevent import (
ImageGenCompletedEvent,
ImageGenCompletedEventTypedDict,
)
from .imagegenpartialimageevent import (
ImageGenPartialImageEvent,
ImageGenPartialImageEventTypedDict,
)
from .imagegenstreamerrorevent import (
ImageGenStreamErrorEvent,
ImageGenStreamErrorEventTypedDict,
)
from openrouter.types import BaseModel
from openrouter.utils import get_discriminator
from pydantic import Discriminator, Tag
from typing import Union
from typing_extensions import Annotated, TypeAliasType, TypedDict
ImageStreamingResponseDataTypedDict = TypeAliasType(
"ImageStreamingResponseDataTypedDict",
Union[
ImageGenStreamErrorEventTypedDict,
ImageGenPartialImageEventTypedDict,
ImageGenCompletedEventTypedDict,
],
)
ImageStreamingResponseData = Annotated[
Union[
Annotated[ImageGenPartialImageEvent, Tag("image_generation.partial_image")],
Annotated[ImageGenCompletedEvent, Tag("image_generation.completed")],
Annotated[ImageGenStreamErrorEvent, Tag("error")],
],
Discriminator(lambda m: get_discriminator(m, "type", "type")),
]
class ImageStreamingResponseTypedDict(TypedDict):
data: ImageStreamingResponseDataTypedDict
class ImageStreamingResponse(BaseModel):
data: ImageStreamingResponseData
+44 -32
View File
@@ -84,6 +84,10 @@ from .outputfunctioncallitem import (
OutputFunctionCallItem,
OutputFunctionCallItemTypedDict,
)
from .outputfusionservertoolitem import (
OutputFusionServerToolItem,
OutputFusionServerToolItemTypedDict,
)
from .outputimagegenerationcallitem import (
OutputImageGenerationCallItem,
OutputImageGenerationCallItemTypedDict,
@@ -104,6 +108,10 @@ from .outputsearchmodelsservertoolitem import (
OutputSearchModelsServerToolItem,
OutputSearchModelsServerToolItemTypedDict,
)
from .outputsubagentservertoolitem import (
OutputSubagentServerToolItem,
OutputSubagentServerToolItemTypedDict,
)
from .outputtexteditorservertoolitem import (
OutputTextEditorServerToolItem,
OutputTextEditorServerToolItemTypedDict,
@@ -388,44 +396,46 @@ InputsUnion1TypedDict = TypeAliasType(
OutputFileSearchCallItemTypedDict,
EasyInputMessageTypedDict,
InputMessageItemTypedDict,
CustomToolCallOutputItemTypedDict,
LocalShellCallOutputItemTypedDict,
OutputToolSearchServerToolItemTypedDict,
OutputFileSearchServerToolItemTypedDict,
OutputWebSearchServerToolItemTypedDict,
CustomToolCallOutputItemTypedDict,
OutputImageGenerationCallItemTypedDict,
OutputWebSearchCallItemTypedDict,
OutputDatetimeItemTypedDict,
OutputMcpServerToolItemTypedDict,
FunctionCallOutputItemTypedDict,
LocalShellCallItemTypedDict,
OutputSearchModelsServerToolItemTypedDict,
FunctionCallOutputItemTypedDict,
ApplyPatchCallItemTypedDict,
OutputDatetimeItemTypedDict,
McpApprovalResponseItemTypedDict,
McpApprovalRequestItemTypedDict,
McpListToolsItemTypedDict,
ApplyPatchCallOutputItemTypedDict,
McpApprovalResponseItemTypedDict,
OutputBrowserUseServerToolItemTypedDict,
McpApprovalRequestItemTypedDict,
OutputMcpServerToolItemTypedDict,
OutputTextEditorServerToolItemTypedDict,
OutputApplyPatchServerToolItemTypedDict,
CustomToolCallItemTypedDict,
ShellCallItemTypedDict,
InputsMessageTypedDict,
OutputMemoryServerToolItemTypedDict,
OutputCustomToolCallItemTypedDict,
ShellCallItemTypedDict,
ShellCallOutputItemTypedDict,
OutputComputerCallItemTypedDict,
OutputCodeInterpreterCallItemTypedDict,
OutputImageGenerationServerToolItemTypedDict,
OutputBashServerToolItemTypedDict,
OutputCustomToolCallItemTypedDict,
OutputComputerCallItemTypedDict,
CustomToolCallItemTypedDict,
ShellCallOutputItemTypedDict,
McpCallItemTypedDict,
OutputImageGenerationServerToolItemTypedDict,
OutputFunctionCallItemTypedDict,
OutputBashServerToolItemTypedDict,
FunctionCallItemTypedDict,
OutputAdvisorServerToolItemTypedDict,
OutputWebFetchServerToolItemTypedDict,
ReasoningItemTypedDict,
OutputCodeInterpreterServerToolItemTypedDict,
OutputSubagentServerToolItemTypedDict,
InputsReasoningTypedDict,
OutputCodeInterpreterServerToolItemTypedDict,
OutputWebFetchServerToolItemTypedDict,
OutputAdvisorServerToolItemTypedDict,
OutputFusionServerToolItemTypedDict,
],
)
@@ -438,44 +448,46 @@ InputsUnion1 = TypeAliasType(
OutputFileSearchCallItem,
EasyInputMessage,
InputMessageItem,
CustomToolCallOutputItem,
LocalShellCallOutputItem,
OutputToolSearchServerToolItem,
OutputFileSearchServerToolItem,
OutputWebSearchServerToolItem,
CustomToolCallOutputItem,
OutputImageGenerationCallItem,
OutputWebSearchCallItem,
OutputDatetimeItem,
OutputMcpServerToolItem,
FunctionCallOutputItem,
LocalShellCallItem,
OutputSearchModelsServerToolItem,
FunctionCallOutputItem,
ApplyPatchCallItem,
OutputDatetimeItem,
McpApprovalResponseItem,
McpApprovalRequestItem,
McpListToolsItem,
ApplyPatchCallOutputItem,
McpApprovalResponseItem,
OutputBrowserUseServerToolItem,
McpApprovalRequestItem,
OutputMcpServerToolItem,
OutputTextEditorServerToolItem,
OutputApplyPatchServerToolItem,
CustomToolCallItem,
ShellCallItem,
InputsMessage,
OutputMemoryServerToolItem,
OutputCustomToolCallItem,
ShellCallItem,
ShellCallOutputItem,
OutputComputerCallItem,
OutputCodeInterpreterCallItem,
OutputImageGenerationServerToolItem,
OutputBashServerToolItem,
OutputCustomToolCallItem,
OutputComputerCallItem,
CustomToolCallItem,
ShellCallOutputItem,
McpCallItem,
OutputImageGenerationServerToolItem,
OutputFunctionCallItem,
OutputBashServerToolItem,
FunctionCallItem,
OutputAdvisorServerToolItem,
OutputWebFetchServerToolItem,
ReasoningItem,
OutputCodeInterpreterServerToolItem,
OutputSubagentServerToolItem,
InputsReasoning,
OutputCodeInterpreterServerToolItem,
OutputWebFetchServerToolItem,
OutputAdvisorServerToolItem,
OutputFusionServerToolItem,
],
)
@@ -27,10 +27,10 @@ class LegacyWebSearchServerToolTypedDict(TypedDict):
type: LegacyWebSearchServerToolType
engine: NotRequired[WebSearchEngineEnum]
r"""Which search engine to use. \"auto\" (default) uses native if the provider supports it, otherwise Exa. \"native\" forces the provider's built-in search. \"exa\" forces the Exa search API. \"firecrawl\" uses Firecrawl (requires BYOK). \"parallel\" uses the Parallel search API."""
r"""Which search engine to use. \"auto\" (default) uses native if the provider supports it, otherwise Exa. \"native\" forces the provider's built-in search. \"exa\" forces the Exa search API. \"firecrawl\" uses Firecrawl (requires BYOK). \"parallel\" uses the Parallel search API. \"perplexity\" uses the Perplexity Search API (raw ranked results)."""
filters: NotRequired[Nullable[WebSearchDomainFilterTypedDict]]
max_results: NotRequired[int]
r"""Maximum number of search results to return per search call. Defaults to 5. Applies to Exa, Firecrawl, and Parallel engines; ignored with native provider search."""
r"""Maximum number of search results to return per search call. Defaults to 5. Applies to Exa, Firecrawl, Parallel, and Perplexity engines; ignored with native provider search. Perplexity supports a maximum of 20; values above 20 are clamped."""
search_context_size: NotRequired[SearchContextSizeEnum]
r"""Size of the search context for web search tools"""
user_location: NotRequired[Nullable[WebSearchUserLocationTypedDict]]
@@ -45,12 +45,12 @@ class LegacyWebSearchServerTool(BaseModel):
engine: Annotated[
Optional[WebSearchEngineEnum], PlainValidator(validate_open_enum(False))
] = None
r"""Which search engine to use. \"auto\" (default) uses native if the provider supports it, otherwise Exa. \"native\" forces the provider's built-in search. \"exa\" forces the Exa search API. \"firecrawl\" uses Firecrawl (requires BYOK). \"parallel\" uses the Parallel search API."""
r"""Which search engine to use. \"auto\" (default) uses native if the provider supports it, otherwise Exa. \"native\" forces the provider's built-in search. \"exa\" forces the Exa search API. \"firecrawl\" uses Firecrawl (requires BYOK). \"parallel\" uses the Parallel search API. \"perplexity\" uses the Perplexity Search API (raw ranked results)."""
filters: OptionalNullable[WebSearchDomainFilter] = UNSET
max_results: Optional[int] = None
r"""Maximum number of search results to return per search call. Defaults to 5. Applies to Exa, Firecrawl, and Parallel engines; ignored with native provider search."""
r"""Maximum number of search results to return per search call. Defaults to 5. Applies to Exa, Firecrawl, Parallel, and Perplexity engines; ignored with native provider search. Perplexity supports a maximum of 20; values above 20 are clamped."""
search_context_size: Annotated[
Optional[SearchContextSizeEnum], PlainValidator(validate_open_enum(False))
@@ -0,0 +1,17 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .workspacebudget import WorkspaceBudget, WorkspaceBudgetTypedDict
from openrouter.types import BaseModel
from typing import List
from typing_extensions import TypedDict
class ListWorkspaceBudgetsResponseTypedDict(TypedDict):
data: List[WorkspaceBudgetTypedDict]
r"""List of budgets configured for the workspace"""
class ListWorkspaceBudgetsResponse(BaseModel):
data: List[WorkspaceBudget]
r"""List of budgets configured for the workspace"""
@@ -0,0 +1,33 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from openrouter.types import BaseModel, Nullable
import pydantic
from pydantic import ConfigDict
from typing import Any, Dict
from typing_extensions import TypedDict
class MessagesFallbackParamTypedDict(TypedDict):
r"""Fallback model to try when the primary model fails or refuses. Only the `model` field is supported; per-attempt overrides are rejected."""
model: str
class MessagesFallbackParam(BaseModel):
r"""Fallback model to try when the primary model fails or refuses. Only the `model` field is supported; per-attempt overrides are rejected."""
model_config = ConfigDict(
populate_by_name=True, arbitrary_types_allowed=True, extra="allow"
)
__pydantic_extra__: Dict[str, Nullable[Any]] = pydantic.Field(init=False)
model: str
@property
def additional_properties(self):
return self.__pydantic_extra__
@additional_properties.setter
def additional_properties(self, value):
self.__pydantic_extra__ = value # pyright: ignore[reportIncompatibleVariableOverride]
+18 -5
View File
@@ -50,6 +50,7 @@ from .imagegenerationservertool_openrouter import (
ImageGenerationServerToolOpenRouter,
ImageGenerationServerToolOpenRouterTypedDict,
)
from .messagesfallbackparam import MessagesFallbackParam, MessagesFallbackParamTypedDict
from .messagesmessageparam import MessagesMessageParam, MessagesMessageParamTypedDict
from .messagesoutputconfig import MessagesOutputConfig, MessagesOutputConfigTypedDict
from .moderationplugin import ModerationPlugin, ModerationPluginTypedDict
@@ -297,11 +298,11 @@ class ContextManagement(BaseModel):
edits: Optional[List[Edit]] = None
class MetadataTypedDict(TypedDict):
class MessagesRequestMetadataTypedDict(TypedDict):
user_id: NotRequired[Nullable[str]]
class Metadata(BaseModel):
class MessagesRequestMetadata(BaseModel):
user_id: OptionalNullable[str] = UNSET
@model_serializer(mode="wrap")
@@ -1046,8 +1047,10 @@ class MessagesRequestTypedDict(TypedDict):
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."""
context_management: NotRequired[Nullable[ContextManagementTypedDict]]
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."""
max_tokens: NotRequired[int]
metadata: NotRequired[MetadataTypedDict]
metadata: NotRequired[MessagesRequestMetadataTypedDict]
models: NotRequired[List[str]]
output_config: NotRequired[MessagesOutputConfigTypedDict]
r"""Configuration for controlling output behavior. Supports the effort parameter and structured output format."""
@@ -1088,9 +1091,12 @@ class MessagesRequest(BaseModel):
context_management: OptionalNullable[ContextManagement] = UNSET
fallbacks: OptionalNullable[List[MessagesFallbackParam]] = UNSET
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."""
max_tokens: Optional[int] = None
metadata: Optional[Metadata] = None
metadata: Optional[MessagesRequestMetadata] = None
models: Optional[List[str]] = None
@@ -1144,6 +1150,7 @@ class MessagesRequest(BaseModel):
optional_fields = [
"cache_control",
"context_management",
"fallbacks",
"max_tokens",
"metadata",
"models",
@@ -1166,7 +1173,13 @@ class MessagesRequest(BaseModel):
"trace",
"user",
]
nullable_fields = ["context_management", "messages", "provider", "speed"]
nullable_fields = [
"context_management",
"fallbacks",
"messages",
"provider",
"speed",
]
null_default_fields = []
serialized = handler(self)
+14
View File
@@ -3,7 +3,9 @@
from __future__ import annotations
from .defaultparameters import DefaultParameters, DefaultParametersTypedDict
from .modelarchitecture import ModelArchitecture, ModelArchitectureTypedDict
from .modelbenchmarks import ModelBenchmarks, ModelBenchmarksTypedDict
from .modellinks import ModelLinks, ModelLinksTypedDict
from .modelreasoning import ModelReasoning, ModelReasoningTypedDict
from .parameter import Parameter
from .perrequestlimits import PerRequestLimits, PerRequestLimitsTypedDict
from .publicpricing import PublicPricing, PublicPricingTypedDict
@@ -51,6 +53,8 @@ class ModelTypedDict(TypedDict):
r"""List of supported voice identifiers for TTS models. Null for non-TTS models."""
top_provider: TopProviderInfoTypedDict
r"""Information about the top provider for this model"""
benchmarks: NotRequired[ModelBenchmarksTypedDict]
r"""Third-party benchmark rankings for this model. Omitted when no benchmark data is available."""
description: NotRequired[str]
r"""Description of the model"""
expiration_date: NotRequired[Nullable[str]]
@@ -59,6 +63,8 @@ class ModelTypedDict(TypedDict):
r"""Hugging Face model identifier, if applicable"""
knowledge_cutoff: NotRequired[Nullable[str]]
r"""The date up to which the model was trained on data. ISO 8601 date string (YYYY-MM-DD) or null if unknown."""
reasoning: NotRequired[ModelReasoningTypedDict]
r"""Reasoning effort configuration. Omitted for non-reasoning models and dynamic router models."""
class Model(BaseModel):
@@ -105,6 +111,9 @@ class Model(BaseModel):
top_provider: TopProviderInfo
r"""Information about the top provider for this model"""
benchmarks: Optional[ModelBenchmarks] = None
r"""Third-party benchmark rankings for this model. Omitted when no benchmark data is available."""
description: Optional[str] = None
r"""Description of the model"""
@@ -117,13 +126,18 @@ class Model(BaseModel):
knowledge_cutoff: OptionalNullable[str] = UNSET
r"""The date up to which the model was trained on data. ISO 8601 date string (YYYY-MM-DD) or null if unknown."""
reasoning: Optional[ModelReasoning] = None
r"""Reasoning effort configuration. Omitted for non-reasoning models and dynamic router models."""
@model_serializer(mode="wrap")
def serialize_model(self, handler):
optional_fields = [
"benchmarks",
"description",
"expiration_date",
"hugging_face_id",
"knowledge_cutoff",
"reasoning",
]
nullable_fields = [
"context_length",
@@ -0,0 +1,27 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .aabenchmarkentry import AABenchmarkEntry, AABenchmarkEntryTypedDict
from .dabenchmarkentry import DABenchmarkEntry, DABenchmarkEntryTypedDict
from openrouter.types import BaseModel
from typing import List, Optional
from typing_extensions import NotRequired, TypedDict
class ModelBenchmarksTypedDict(TypedDict):
r"""Third-party benchmark rankings for this model. Omitted when no benchmark data is available."""
design_arena: List[DABenchmarkEntryTypedDict]
r"""Design Arena ELO rankings across arena+category pairs."""
artificial_analysis: NotRequired[AABenchmarkEntryTypedDict]
r"""Artificial Analysis benchmark index scores."""
class ModelBenchmarks(BaseModel):
r"""Third-party benchmark rankings for this model. Omitted when no benchmark data is available."""
design_arena: List[DABenchmarkEntry]
r"""Design Arena ELO rankings across arena+category pairs."""
artificial_analysis: Optional[AABenchmarkEntry] = None
r"""Artificial Analysis benchmark index scores."""
+107
View File
@@ -0,0 +1,107 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .reasoningeffort import ReasoningEffort
from openrouter.types import (
BaseModel,
Nullable,
OptionalNullable,
UNSET,
UNSET_SENTINEL,
UnrecognizedStr,
)
from openrouter.utils import validate_open_enum
from pydantic import model_serializer
from pydantic.functional_validators import PlainValidator
from typing import List, Literal, Optional, Union
from typing_extensions import Annotated, NotRequired, TypedDict
DefaultEffort = Union[
Literal[
"max",
"xhigh",
"high",
"medium",
"low",
"minimal",
"none",
],
UnrecognizedStr,
]
r"""Default reasoning effort when the client enables reasoning without specifying effort. Maps to `reasoning.effort` in chat requests. When `\"none\"`, prefer omitting effort unless the user explicitly disables reasoning."""
class ModelReasoningTypedDict(TypedDict):
r"""Reasoning effort configuration. Omitted for non-reasoning models and dynamic router models."""
mandatory: bool
r"""When true, reasoning cannot be disabled and effort \"none\" is rejected."""
default_effort: NotRequired[Nullable[DefaultEffort]]
default_enabled: NotRequired[bool]
r"""Default reasoning enabled state when the client does not set `reasoning.enabled`."""
supported_efforts: NotRequired[Nullable[List[Nullable[ReasoningEffort]]]]
r"""Allowed reasoning effort values for this model, in descending effort order (highest first). Null means no allowlist — all gateway effort values are accepted."""
supports_max_tokens: NotRequired[bool]
r"""Present and `true` when the model accepts `reasoning.max_tokens` in requests (Anthropic-style) instead of or in addition to `reasoning.effort`. Omitted otherwise."""
class ModelReasoning(BaseModel):
r"""Reasoning effort configuration. Omitted for non-reasoning models and dynamic router models."""
mandatory: bool
r"""When true, reasoning cannot be disabled and effort \"none\" is rejected."""
default_effort: Annotated[
OptionalNullable[DefaultEffort], PlainValidator(validate_open_enum(False))
] = UNSET
default_enabled: Optional[bool] = None
r"""Default reasoning enabled state when the client does not set `reasoning.enabled`."""
supported_efforts: OptionalNullable[
List[
Annotated[
Nullable[ReasoningEffort], PlainValidator(validate_open_enum(False))
]
]
] = UNSET
r"""Allowed reasoning effort values for this model, in descending effort order (highest first). Null means no allowlist — all gateway effort values are accepted."""
supports_max_tokens: Optional[bool] = None
r"""Present and `true` when the model accepts `reasoning.max_tokens` in requests (Anthropic-style) instead of or in addition to `reasoning.effort`. Omitted otherwise."""
@model_serializer(mode="wrap")
def serialize_model(self, handler):
optional_fields = [
"default_effort",
"default_enabled",
"supported_efforts",
"supports_max_tokens",
]
nullable_fields = ["default_effort", "supported_efforts"]
null_default_fields = []
serialized = handler(self)
m = {}
for n, f in type(self).model_fields.items():
k = f.alias or n
val = serialized.get(k)
serialized.pop(k, None)
optional_nullable = k in optional_fields and k in nullable_fields
is_set = (
self.__pydantic_fields_set__.intersection({n})
or k in null_default_fields
) # pylint: disable=no-member
if val is not None and val != UNSET_SENTINEL:
m[k] = val
elif val != UNSET_SENTINEL and (
not k in optional_fields or (optional_nullable and is_set)
):
m[k] = val
return m
@@ -0,0 +1,20 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .model import Model, ModelTypedDict
from openrouter.types import BaseModel
from typing_extensions import TypedDict
class ModelResponseTypedDict(TypedDict):
r"""Single model response"""
data: ModelTypedDict
r"""Information about an AI model available on OpenRouter"""
class ModelResponse(BaseModel):
r"""Single model response"""
data: Model
r"""Information about an AI model available on OpenRouter"""
@@ -15,6 +15,7 @@ from typing_extensions import Annotated, NotRequired, TypedDict
class ObservabilityArizeDestinationConfigTypedDict(TypedDict):
api_key: str
model_id: str
r"""The name of the tracing project in Arize AX"""
space_key: str
base_url: NotRequired[str]
headers: NotRequired[Dict[str, str]]
@@ -25,6 +26,7 @@ class ObservabilityArizeDestinationConfig(BaseModel):
api_key: Annotated[str, pydantic.Field(alias="apiKey")]
model_id: Annotated[str, pydantic.Field(alias="modelId")]
r"""The name of the tracing project in Arize AX"""
space_key: Annotated[str, pydantic.Field(alias="spaceKey")]
@@ -1,6 +1,7 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .apierrortype import APIErrorType
from .applypatchservertool import ApplyPatchServerTool, ApplyPatchServerToolTypedDict
from .baseinputs_union import BaseInputsUnion, BaseInputsUnionTypedDict
from .basereasoningconfig import BaseReasoningConfig, BaseReasoningConfigTypedDict
@@ -203,6 +204,8 @@ class OpenResponsesResultTypedDict(TypedDict):
usage: NotRequired[Nullable[UsageTypedDict]]
r"""Token usage information for the response"""
user: NotRequired[Nullable[str]]
error_type: NotRequired[APIErrorType]
r"""Canonical OpenRouter error type, stable across all API formats"""
openrouter_metadata: NotRequired[OpenRouterMetadataTypedDict]
@@ -285,6 +288,11 @@ class OpenResponsesResult(BaseModel):
user: OptionalNullable[str] = UNSET
error_type: Annotated[
Optional[APIErrorType], PlainValidator(validate_open_enum(False))
] = None
r"""Canonical OpenRouter error type, stable across all API formats"""
openrouter_metadata: Optional[OpenRouterMetadata] = None
@model_serializer(mode="wrap")
@@ -306,6 +314,7 @@ class OpenResponsesResult(BaseModel):
"truncation",
"usage",
"user",
"error_type",
"openrouter_metadata",
]
nullable_fields = [
@@ -2,6 +2,7 @@
from __future__ import annotations
from .fusionanalysisresult import FusionAnalysisResult, FusionAnalysisResultTypedDict
from .fusionsource import FusionSource, FusionSourceTypedDict
from .toolcallstatus import ToolCallStatus
from openrouter.types import BaseModel
from openrouter.utils import validate_open_enum
@@ -60,6 +61,8 @@ class OutputFusionServerToolItemTypedDict(TypedDict):
id: NotRequired[str]
responses: NotRequired[List[ResponseTypedDict]]
r"""Analysis models that produced a response in this fusion run, with each model's full panel content."""
sources: NotRequired[List[FusionSourceTypedDict]]
r"""Web pages the analysis panels and judge retrieved via web search during this fusion run, deduplicated by URL across the whole run. Present when at least one model cited a source."""
class OutputFusionServerToolItem(BaseModel):
@@ -85,3 +88,6 @@ class OutputFusionServerToolItem(BaseModel):
responses: Optional[List[Response]] = None
r"""Analysis models that produced a response in this fusion run, with each model's full panel content."""
sources: Optional[List[FusionSource]] = None
r"""Web pages the analysis panels and judge retrieved via web search during this fusion run, deduplicated by URL across the whole run. Present when at least one model cited a source."""
+7 -1
View File
@@ -81,6 +81,10 @@ from .outputshellcalloutputitem import (
OutputShellCallOutputItem,
OutputShellCallOutputItemTypedDict,
)
from .outputsubagentservertoolitem import (
OutputSubagentServerToolItem,
OutputSubagentServerToolItemTypedDict,
)
from .outputtexteditorservertoolitem import (
OutputTextEditorServerToolItem,
OutputTextEditorServerToolItemTypedDict,
@@ -136,8 +140,9 @@ OutputItemsTypedDict = TypeAliasType(
OutputWebFetchServerToolItemTypedDict,
OutputCodeInterpreterServerToolItemTypedDict,
OutputReasoningItemTypedDict,
OutputFusionServerToolItemTypedDict,
OutputAdvisorServerToolItemTypedDict,
OutputSubagentServerToolItemTypedDict,
OutputFusionServerToolItemTypedDict,
],
)
r"""An output item from the response"""
@@ -172,6 +177,7 @@ OutputItems = Annotated[
],
Annotated[OutputMcpServerToolItem, Tag("openrouter:mcp")],
Annotated[OutputMemoryServerToolItem, Tag("openrouter:memory")],
Annotated[OutputSubagentServerToolItem, Tag("openrouter:subagent")],
Annotated[OutputTextEditorServerToolItem, Tag("openrouter:text_editor")],
Annotated[OutputToolSearchServerToolItem, Tag("openrouter:tool_search")],
Annotated[OutputWebFetchServerToolItem, Tag("openrouter:web_fetch")],
@@ -0,0 +1,55 @@
"""Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT."""
from __future__ import annotations
from .toolcallstatus import ToolCallStatus
from openrouter.types import BaseModel
from openrouter.utils import validate_open_enum
from pydantic.functional_validators import PlainValidator
from typing import Literal, Optional
from typing_extensions import Annotated, NotRequired, TypedDict
OutputSubagentServerToolItemType = Literal["openrouter:subagent",]
class OutputSubagentServerToolItemTypedDict(TypedDict):
r"""An openrouter:subagent server tool output item"""
status: ToolCallStatus
type: OutputSubagentServerToolItemType
error: NotRequired[str]
r"""Error message when the subagent task did not produce an outcome."""
id: NotRequired[str]
model: NotRequired[str]
r"""Slug of the worker model that executed the task."""
outcome: NotRequired[str]
r"""The worker model's result (the outcome text returned to the delegating model)."""
task_description: NotRequired[str]
r"""The task description the delegating model sent to the worker."""
task_name: NotRequired[str]
r"""The short task identifier the delegating model supplied."""
class OutputSubagentServerToolItem(BaseModel):
r"""An openrouter:subagent server tool output item"""
status: Annotated[ToolCallStatus, PlainValidator(validate_open_enum(False))]
type: OutputSubagentServerToolItemType
error: Optional[str] = None
r"""Error message when the subagent task did not produce an outcome."""
id: Optional[str] = None
model: Optional[str] = None
r"""Slug of the worker model that executed the task."""
outcome: Optional[str] = None
r"""The worker model's result (the outcome text returned to the delegating model)."""
task_description: Optional[str] = None
r"""The task description the delegating model sent to the worker."""
task_name: Optional[str] = None
r"""The short task identifier the delegating model supplied."""

Some files were not shown because too many files have changed in this diff Show More