Whisper STT: Add Toggle for "[Speaking]" Prompt

public/i18n/en/translation.json

@@ -84,6 +84,7 @@
     "whisper_vad_silence": "VAD Silence",
     "whisper_silence_note": "Duration in ms to wait for to detect end of speech segment. Default: 240",
     "whisper_use_gpu": "GPU Acceleration",
+    "whisper_speaking_prompt": "Toggle Speaking Prompt",
 
     "deepgram_title": "Deepgram",
     "deepgram_key": "Key",

public/i18n/zh_cn/translation.json

@@ -76,6 +76,15 @@
     "speechly_title": "Speechly",
     "speechly_appid": "App ID",
 
+    "whisper_title": "Whisper",
+    "whisper_model_path": "Whisper 模型",
+    "whisper_translate_to_english": "翻译为英文",
+    "whisper_lang_desc_note": "注意：质量因选择的语言和模型而异",
+    "whisper_vad_silence": "VAD 静音",
+    "whisper_silence_note": "等待检测语音段落结尾的毫秒数。默认值：240",
+    "whisper_use_gpu": "GPU 加速",
+    "whisper_speaking_prompt": "切换说话提示",
+
     "deepgram_title": "Deepgram",
     "deepgram_key": "Key",
     "deepgram_quality": "质量",

public/i18n/zh_tw/translation.json

@@ -76,6 +76,15 @@
     "speechly_title": "Speechly",
     "speechly_appid": "App ID",
 
+    "whisper_title": "Whisper",
+    "whisper_model_path": "Whisper 模型",
+    "whisper_translate_to_english": "翻譯為英文",
+    "whisper_lang_desc_note": "注意：質量因選擇的語言和模型而異",
+    "whisper_vad_silence": "VAD 靜音",
+    "whisper_silence_note": "等待檢測語音段落結尾的毫秒數。默認值：240",
+    "whisper_use_gpu": "GPU 加速",
+    "whisper_speaking_prompt": "切換說話提示",
+
     "deepgram_title": "Deepgram",
     "deepgram_key": "Key",
     "deepgram_quality": "質量",

src-tauri/src/services/whisper_stt/mod.rs

@@ -65,6 +65,7 @@ pub struct WhisperArgs {
     translate_to_english: bool,
     silence_interval: u64,
     use_gpu: bool,
+    show_speaking_prompt: bool,
 }
 
 pub fn init<R: Runtime>() -> plugin::TauriPlugin<R> {
@@ -200,7 +201,8 @@ pub async fn start<R: Runtime>(app: AppHandle<R>, args: WhisperArgs) -> Result<(
                 event = activity_rx.next() => {
                     match event {
                         Some(VadActivity::SpeechStart) => {
-                            if app.emit("whisper_stt_interim", "[speaking]").is_err() {
+                            let interim_text = if args.show_speaking_prompt { "[speaking]" } else { "" };
+                            if app.emit("whisper_stt_interim", interim_text).is_err() {
                                 eprintln!("wasn't able to emit to frontend {}:{}", file!(), line!());
                             }
                         },

src/server/services/stt/schema.ts

@@ -49,6 +49,7 @@ export const Service_STT_Schema = z.object({
     translateToEnglish: zSafe(z.coerce.boolean(), false),
     silenceInterval: zSafe(zStringNumber(), "240"),
     useGpu: zSafe(z.coerce.boolean(), true),
+    showSpeakingPrompt: zSafe(z.coerce.boolean(), true),
   }).default({}),
   deepgram: z.object({
     device: zSafe(z.coerce.string(), "default"),

src/server/services/stt/services/whisper.ts

@@ -43,6 +43,7 @@ export class STT_WhisperService implements ISTTService {
             translateToEnglish: this.state.translateToEnglish,
             silenceInterval: parseInt(this.state.silenceInterval),
             useGpu: this.state.useGpu,
+            showSpeakingPrompt: this.state.showSpeakingPrompt,
         },
     }).catch(err => {
         this.#initialized = false;

src/server/ui/inspector/inspector_stt.tsx

@@ -190,6 +190,7 @@ const Whisper: FC = () => {
     <InputRange label="stt.whisper_vad_silence" step="30" min="120" max="2000" value={data.silenceInterval} onChange={e => handleUpdate("silenceInterval", e.target.value)} />
     <Inspector.Description>{t('stt.whisper_silence_note')}</Inspector.Description>
     <InputCheckbox label="stt.whisper_use_gpu" onChange={e => handleUpdate("useGpu", e)} value={data.useGpu}/>
+    <InputCheckbox label="stt.whisper_speaking_prompt" onChange={e => handleUpdate("showSpeakingPrompt", e)} value={data.showSpeakingPrompt}/>
     <InputFilePath
       label="stt.whisper_model_path"
       value={data.modelPath}