diff --git a/src/_locales/en/main.json b/src/_locales/en/main.json
index c22e405..baa0dcc 100644
--- a/src/_locales/en/main.json
+++ b/src/_locales/en/main.json
@@ -101,5 +101,6 @@
   "Store to Independent Conversation Page": "Store to Independent Conversation Page",
   "Keep Conversation Window in Background": "Keep conversation window in background, so that you can use shortcut keys to call it up in any program",
   "Max Response Token Length": "Max Response Token Length",
+  "Max Conversation Length": "Max Conversation Length",
   "Always pin the floating window": "Always pin the floating window"
 }
diff --git a/src/_locales/zh-hans/main.json b/src/_locales/zh-hans/main.json
index b78a99e..172851e 100644
--- a/src/_locales/zh-hans/main.json
+++ b/src/_locales/zh-hans/main.json
@@ -101,5 +101,6 @@
   "Store to Independent Conversation Page": "收纳到独立对话页",
   "Keep Conversation Window in Background": "保持对话窗口在后台, 以便在任何程序中使用快捷键呼出",
   "Max Response Token Length": "响应的最大token长度",
+  "Max Conversation Length": "对话处理的最大长度",
   "Always pin the floating window": "总是固定浮动窗口"
 }
diff --git a/src/_locales/zh-hant/main.json b/src/_locales/zh-hant/main.json
index 85282c5..96585f0 100644
--- a/src/_locales/zh-hant/main.json
+++ b/src/_locales/zh-hant/main.json
@@ -101,5 +101,6 @@
   "Store to Independent Conversation Page": "收納到獨立對話頁",
   "Keep Conversation Window in Background": "保持對話窗口在後臺, 以便在任何程序中使用快捷鍵呼出",
   "Max Response Token Length": "響應的最大token長度",
+  "Max Conversation Length": "對話處理的最大長度",
   "Always pin the floating window": "總是固定浮動視窗"
 }
diff --git a/src/background/apis/azure-openai-api.mjs b/src/background/apis/azure-openai-api.mjs
index a26aa1b..776efc4 100644
--- a/src/background/apis/azure-openai-api.mjs
+++ b/src/background/apis/azure-openai-api.mjs
@@ -13,7 +13,10 @@ export async function generateAnswersWithAzureOpenaiApi(port, question, session)
   const { controller, messageListener } = setAbortController(port)
   const config = await getUserConfig()
 
-  const prompt = getConversationPairs(session.conversationRecords, false)
+  const prompt = getConversationPairs(
+    session.conversationRecords.slice(-config.maxConversationContextLength),
+    false,
+  )
   prompt.unshift({ role: 'system', content: await getChatSystemPromptBase() })
   prompt.push({ role: 'user', content: question })
 
diff --git a/src/background/apis/custom-api.mjs b/src/background/apis/custom-api.mjs
index 157f32b..7187fda 100644
--- a/src/background/apis/custom-api.mjs
+++ b/src/background/apis/custom-api.mjs
@@ -21,10 +21,13 @@ import { getCustomApiPromptBase, pushRecord, setAbortController } from './shared
 export async function generateAnswersWithCustomApi(port, question, session, apiKey, modelName) {
   const { controller, messageListener } = setAbortController(port)
 
-  const prompt = getConversationPairs(session.conversationRecords, false)
+  const config = await getUserConfig()
+  const prompt = getConversationPairs(
+    session.conversationRecords.slice(-config.maxConversationContextLength),
+    false,
+  )
   prompt.unshift({ role: 'system', content: await getCustomApiPromptBase() })
   prompt.push({ role: 'user', content: question })
-  const config = await getUserConfig()
   const apiUrl = config.customModelApiUrl
 
   let answer = ''
diff --git a/src/background/apis/openai-api.mjs b/src/background/apis/openai-api.mjs
index beb6882..22ed1f2 100644
--- a/src/background/apis/openai-api.mjs
+++ b/src/background/apis/openai-api.mjs
@@ -27,11 +27,14 @@ export async function generateAnswersWithGptCompletionApi(
 ) {
   const { controller, messageListener } = setAbortController(port)
 
+  const config = await getUserConfig()
   const prompt =
     (await getCompletionPromptBase()) +
-    getConversationPairs(session.conversationRecords, true) +
+    getConversationPairs(
+      session.conversationRecords.slice(-config.maxConversationContextLength),
+      true,
+    ) +
     `Human: ${question}\nAI: `
-  const config = await getUserConfig()
   const apiUrl = config.customOpenAiApiUrl
 
   let answer = ''
@@ -89,10 +92,13 @@ export async function generateAnswersWithGptCompletionApi(
 export async function generateAnswersWithChatgptApi(port, question, session, apiKey, modelName) {
   const { controller, messageListener } = setAbortController(port)
 
-  const prompt = getConversationPairs(session.conversationRecords, false)
+  const config = await getUserConfig()
+  const prompt = getConversationPairs(
+    session.conversationRecords.slice(-config.maxConversationContextLength),
+    false,
+  )
   prompt.unshift({ role: 'system', content: await getChatSystemPromptBase() })
   prompt.push({ role: 'user', content: question })
-  const config = await getUserConfig()
   const apiUrl = config.customOpenAiApiUrl
 
   let answer = ''
diff --git a/src/config/index.mjs b/src/config/index.mjs
index 1b85dd5..18eadb8 100644
--- a/src/config/index.mjs
+++ b/src/config/index.mjs
@@ -104,6 +104,7 @@ export const defaultConfig = {
   // advanced
 
   maxResponseTokenLength: 1000,
+  maxConversationContextLength: 9,
   customChatGptWebApiUrl: 'https://chat.openai.com',
   customChatGptWebApiPath: '/backend-api/conversation',
   customOpenAiApiUrl: 'https://api.openai.com',
diff --git a/src/popup/Popup.jsx b/src/popup/Popup.jsx
index 841a3ac..eee368f 100644
--- a/src/popup/Popup.jsx
+++ b/src/popup/Popup.jsx
@@ -27,7 +27,14 @@ import wechatpay from './donation/wechatpay.jpg'
 import bugmeacoffee from './donation/bugmeacoffee.png'
 import { useWindowTheme } from '../hooks/use-window-theme.mjs'
 import { languageList } from '../config/language.mjs'
-import { isEdge, isFirefox, isMobile, isSafari, openUrl } from '../utils/index.mjs'
+import {
+  isEdge,
+  isFirefox,
+  isMobile,
+  isSafari,
+  openUrl,
+  parseIntWithClamp,
+} from '../utils/index.mjs'
 import { useTranslation } from 'react-i18next'
 
 function GeneralPart({ config, updateConfig }) {
@@ -413,11 +420,25 @@ function AdvancedPart({ config, updateConfig }) {
           step="100"
           value={config.maxResponseTokenLength}
           onChange={(e) => {
-            const value = parseInt(e.target.value)
+            const value = parseIntWithClamp(e.target.value, 1000, 100, 40000)
             updateConfig({ maxResponseTokenLength: value })
           }}
         />
       </label>
+      <label>
+        {t('Max Conversation Length')}
+        <input
+          type="number"
+          min="0"
+          max="100"
+          step="1"
+          value={config.maxConversationContextLength}
+          onChange={(e) => {
+            const value = parseIntWithClamp(e.target.value, 9, 0, 100)
+            updateConfig({ maxConversationContextLength: value })
+          }}
+        />
+      </label>
       <label>
         {t('Custom ChatGPT Web API Url')}
         <input
diff --git a/src/utils/index.mjs b/src/utils/index.mjs
index de7cda4..6844c1e 100644
--- a/src/utils/index.mjs
+++ b/src/utils/index.mjs
@@ -14,6 +14,7 @@ export * from './is-mobile'
 export * from './is-safari'
 export * from './limited-fetch'
 export * from './open-url'
+export * from './parse-int-with-clamp'
 export * from './set-element-position-in-viewport'
 export * from './stream-async-iterable'
 export * from './update-ref-height'
diff --git a/src/utils/parse-int-with-clamp.js b/src/utils/parse-int-with-clamp.js
new file mode 100644
index 0000000..77c94b5
--- /dev/null
+++ b/src/utils/parse-int-with-clamp.js
@@ -0,0 +1,9 @@
+export function parseIntWithClamp(value, defaultValue, min, max) {
+  value = parseInt(value)
+
+  if (isNaN(value)) value = defaultValue
+  else if (value > max) value = max
+  else if (value < min) value = min
+
+  return value
+}