@goodandready/dsh-voice 0.8.14 → 0.8.16

package/lib/client.js

@@ -123,6 +123,8 @@ window.__ModuleLoader__.load({
       'general': 'General',
       'whisperEndpoint': 'Local whisper: endpoint',
       'whisperEndpointHint': 'POST /inference of a whisper.cpp server',
+      'deepgramEndpoint': 'Deepgram: base URL',
+      'deepgramEndpointHint': 'Base URL for Deepgram or self-hosted deployment (default https://api.deepgram.com)',
       'whisperBin': 'Local whisper: binary',
       'whisperBinHint': 'used when autostart is on',
       'whisperModel': 'Local whisper: model',
@@ -268,6 +270,8 @@ window.__ModuleLoader__.load({
       'general': 'Общее',
       'whisperEndpoint': 'Локальный whisper: endpoint',
       'whisperEndpointHint': 'POST /inference сервера whisper.cpp',
+      'deepgramEndpoint': 'Deepgram: базовый URL',
+      'deepgramEndpointHint': 'Базовый адрес Deepgram или self-hosted развёртывания (по умолчанию https://api.deepgram.com)',
       'whisperBin': 'Локальный whisper: бинарь',
       'whisperBinHint': 'используется при автозапуске',
       'whisperModel': 'Локальный whisper: модель',
@@ -1885,6 +1889,7 @@ window.__ModuleLoader__.load({
           textField('whisperUrl', t('whisperEndpoint'), t('whisperEndpointHint')),
           textField('whisperBin', t('whisperBin'), t('whisperBinHint')),
           textField('whisperModel', t('whisperModel'), t('whisperBinHint')),
+          textField('deepgramBaseUrl', t('deepgramEndpoint'), t('deepgramEndpointHint')),
           React.createElement('label', { className: 'dvs-field' }, t('whisperAutostart'),
             React.createElement('input', {
               type: 'checkbox', checked: !!(draft && draft.autoStart), disabled: !writable,

package/lib/index.js

@@ -97,6 +97,9 @@ export const Config = z.object({
   customProviders: z.array(CustomProvider).default([])
     .description('Own recognition providers, usable in both chains next to the built-in ones.'),
   deepgramKeyEnv: z.string().default('DEEPGRAM_API_KEY'),
+  deepgramBaseUrl: z.string().default('https://api.deepgram.com')
+    .description('Deepgram-compatible host. Point this at a private/self-hosted Deepgram deployment '
+      + 'instead of the public api.deepgram.com endpoint.'),
   groqKeyEnv: z.string().default('GROQ_API_KEY'),
   hfTokenEnv: z.string().default('HF_TOKEN'),
   whisperUrl: z.string().default('http://127.0.0.1:8001/inference'),
@@ -438,6 +441,7 @@ export function apply(ctx, baseConfig) {
         contextGlossary: cfg.contextGlossary !== false,
         sensevoiceUrl: cfg.sensevoiceUrl,
         sensevoiceAutostart: !!cfg.sensevoiceAutostart,
+        deepgramBaseUrl: cfg.deepgramBaseUrl || 'https://api.deepgram.com',
         realtimeStreaming: !!cfg.realtimeStreaming,
         realtimeProvider: cfg.realtimeProvider || 'openai',
         realtimeModel: cfg.realtimeModel || 'gpt-4o-realtime-preview',

package/lib/providers.js

@@ -116,7 +116,8 @@ export function makeProviders(deps, req) {
     const key = await resolveKey(cfg.deepgramKeyEnv)
     if (!key) return { ok: false, provider: 'deepgram', reason: `no ${cfg.deepgramKeyEnv}` }
     const model = pickModel(models, 'deepgram')
-    const url = `https://api.deepgram.com/v1/listen?model=${encodeURIComponent(model)}`
+    const base = String(cfg.deepgramBaseUrl || 'https://api.deepgram.com').replace(/\/+$/, '')
+    const url = `${base}/v1/listen?model=${encodeURIComponent(model)}`
       + (!isAutoLang(lang) ? `&language=${encodeURIComponent(lang)}` : '') + '&smart_format=true'
     const res = await fetchImpl(url, {
       method: 'POST',
@@ -184,7 +185,13 @@ export function makeProviders(deps, req) {
     }
     const form = new FormData()
     form.append('file', new Blob([sendBytes], { type: sendMime }), fileName(sendMime))
-    if (!isAutoLang(lang)) form.append('language', lang)
+    // Unlike OpenAI-compatible APIs, whisper.cpp inherits its server's default
+    // language when the form omits this field (normally `en`). Send `auto`
+    // explicitly so every request detects the spoken language.
+    form.append('language', isAutoLang(lang) ? 'auto' : lang)
+    // Also override a server started with --translate: this plugin always needs
+    // a transcript in the spoken language, never Whisper's English translation.
+    form.append('translate', 'false')
     if (vocab) form.append('prompt', vocab)
     form.append('response_format', 'json')
     const res = await fetchImpl(cfg.whisperUrl, { method: 'POST', body: form, signal })

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@goodandready/dsh-voice",
-  "version": "0.8.14",
+  "version": "0.8.16",
   "description": "Voice input for DeepSeek Harness: dictation chunked by pauses and voice messages, each with its own provider fallback chain (Deepgram, Groq, HuggingFace, local whisper.cpp, plus any OpenAI-compatible API of your own).",
   "license": "MIT",
   "type": "module",
@@ -47,10 +47,7 @@
     },
     "client": {
       "platform": "web",
-      "inject": [
-        "@deepseek-ai/dsh-client-runtime",
-        "@deepseek-ai/dsh-client-ui-slots"
-      ]
+      "inject": []
     }
   },
   "peerDependencies": {
@@ -61,4 +58,4 @@
     "@deepseek-ai/dsh-llm": "^0.1.0-rc.6",
     "@deepseek-ai/schemastery": "^3.18.1"
   }
-}
+}