diff --git a/app/components/chat.tsx b/app/components/chat.tsx
index a52f3b13b..3ce9c993e 100644
--- a/app/components/chat.tsx
+++ b/app/components/chat.tsx
@@ -558,10 +558,12 @@ export function ChatActions(props: {
 
   useEffect(() => {
     if (isFirefox()) config.sttConfig.engine = FIREFOX_DEFAULT_STT_ENGINE;
+    const lang = config.sttConfig.lang;
     setSpeechApi(
       config.sttConfig.engine !== DEFAULT_STT_ENGINE
-        ? new WebTranscriptionApi((transcription) =>
-            onRecognitionEnd(transcription),
+        ? new WebTranscriptionApi(
+            (transcription) => onRecognitionEnd(transcription),
+            lang,
           )
         : new OpenAITranscriptionApi((transcription) =>
             onRecognitionEnd(transcription),
diff --git a/app/components/stt-config.tsx b/app/components/stt-config.tsx
index 478259041..01a63cd5f 100644
--- a/app/components/stt-config.tsx
+++ b/app/components/stt-config.tsx
@@ -55,10 +55,10 @@ export function STTConfigList(props: {
       {props.sttConfig.engine === "Web Speech API" && !isFirefox() && (
         <ListItem title="语言选择">
           <Select
-            value={props.sttConfig.language}
+            value={props.sttConfig.lang}
             onChange={(e) => {
               props.updateConfig(
-                (config) => (config.language = e.currentTarget.value),
+                (config) => (config.lang = e.currentTarget.value),
               );
             }}
           >
diff --git a/app/store/config.ts b/app/store/config.ts
index 01a67d765..b0e4c017a 100644
--- a/app/store/config.ts
+++ b/app/store/config.ts
@@ -95,7 +95,7 @@ export const DEFAULT_CONFIG = {
   sttConfig: {
     enable: true,
     engine: DEFAULT_STT_ENGINE,
-    language: DEFAULT_STT_LANGUAGE,
+    lang: DEFAULT_STT_LANGUAGE,
   },
 };
 
diff --git a/app/utils/speech.ts b/app/utils/speech.ts
index dc8102879..1e69f6031 100644
--- a/app/utils/speech.ts
+++ b/app/utils/speech.ts
@@ -93,7 +93,7 @@ export class WebTranscriptionApi extends SpeechApi {
 
   isListening = () => this.listeningStatus;
 
-  constructor(transcriptionCallback?: TranscriptionCallback) {
+  constructor(transcriptionCallback?: TranscriptionCallback, lang?: string) {
     super();
     if (isFirefox()) return;
     const SpeechRecognition =
@@ -102,7 +102,7 @@ export class WebTranscriptionApi extends SpeechApi {
     this.recognitionInstance = new SpeechRecognition();
     this.recognitionInstance.continuous = true;
     this.recognitionInstance.interimResults = true;
-    this.recognitionInstance.lang = getSTTLang();
+    this.recognitionInstance.lang = lang ?? getSTTLang();
     if (transcriptionCallback) {
       this.onTranscriptionReceived(transcriptionCallback);
     }