stiff

mnemo_cards_admin/src/api/voiceService.ts

@@ -13,13 +13,15 @@ export const voiceServiceApi = {
    * @param voice Voice identifier (optional, defaults to 'alloy')
    * @param language Language code (optional)
    * @param instructions Instructions for speech generation (optional)
+   * @param model Model identifier (optional)
    * @returns Object ID and presigned URL
    */
   async generateTTS(
     text: string,
     voice?: string,
     language?: string,
-    instructions?: string
+    instructions?: string,
+    model?: string
   ): Promise<GenerateTTSResponse> {
     const response = await adminApiClient.post<GenerateTTSResponse>(
       '/api/v2/media/generate-tts',
@@ -28,6 +30,7 @@ export const voiceServiceApi = {
         voice: voice || 'alloy',
         ...(language && { language }),
         ...(instructions && { instructions }),
+        ...(model && { model }),
       }
     )
 

mnemo_cards_admin/src/components/CardVoicesManager.tsx

@@ -43,6 +43,7 @@ export function CardVoicesManager({
   const [dialogVoice, setDialogVoice] = useState('alloy')
   const [dialogLanguage, setDialogLanguage] = useState('es')
   const [dialogInstructions, setDialogInstructions] = useState('pronunciar la frase claramente')
+  const [dialogModel, setDialogModel] = useState('gpt-4o-mini-tts')
 
   // Load voices for the card
   const { data: voicesData, isLoading, refetch } = useQuery({
@@ -124,6 +125,7 @@ export function CardVoicesManager({
     setDialogVoice('alloy')
     setDialogLanguage('es')
     setDialogInstructions('pronunciar la frase claramente')
+    setDialogModel('gpt-4o-mini-tts')
     setShowTTSDialog(true)
   }
 
@@ -146,12 +148,13 @@ export function CardVoicesManager({
       const voice = dialogVoice || 'alloy'
       const language = dialogLanguage || 'es'
       const instructions = dialogInstructions || undefined
+      const model = dialogModel || undefined
       
-      console.log('🔄 CardVoicesManager: Requesting TTS generation', { text, voice, language, instructions })
+      console.log('🔄 CardVoicesManager: Requesting TTS generation', { text, voice, language, instructions, model })
       
       // Generate TTS via backend API
       // Backend calls voice-service (which handles caching) and saves to MinIO
-      const ttsResponse = await voiceServiceApi.generateTTS(text, voice, language, instructions)
+      const ttsResponse = await voiceServiceApi.generateTTS(text, voice, language, instructions, model)
       console.log('✅ CardVoicesManager: TTS generated and saved to MinIO', { 
         objectId: ttsResponse.objectId 
       })
@@ -311,16 +314,27 @@ export function CardVoicesManager({
                 />
               </div>
             </div>
-            <div className="space-y-2">
+            <div className="grid grid-cols-2 gap-4">
-              <Label htmlFor="dialog-instructions">Instructions</Label>
+              <div className="space-y-2">
-              <Textarea
+                <Label htmlFor="dialog-model">Model</Label>
-                id="dialog-instructions"
+                <Input
-                value={dialogInstructions}
+                  id="dialog-model"
-                onChange={(e) => setDialogInstructions(e.target.value)}
+                  value={dialogModel}
-                placeholder="pronunciar la frase claramente"
+                  onChange={(e) => setDialogModel(e.target.value)}
-                className="min-h-[80px]"
+                  placeholder="gpt-4o-mini-tts"
-                disabled={isGeneratingTTS}
+                  disabled={isGeneratingTTS}
-              />
+                />
+              </div>
+              <div className="space-y-2">
+                <Label htmlFor="dialog-instructions">Instructions</Label>
+                <Input
+                  id="dialog-instructions"
+                  value={dialogInstructions}
+                  onChange={(e) => setDialogInstructions(e.target.value)}
+                  placeholder="pronunciar la frase claramente"
+                  disabled={isGeneratingTTS}
+                />
+              </div>
             </div>
           </div>
           <DialogFooter>

mnemo_cards_backend/lib/api/v2/media_api_v2.dart

@@ -511,13 +511,14 @@ class MediaApiV2 {
       final voice = json['voice'] as String?;
       final language = json['language'] as String?;
       final instructions = json['instructions'] as String?;
+      final model = json['model'] as String?;
 
       if (text == null || text.trim().isEmpty) {
         return _badRequest('Text is required');
       }
 
       print(
-        '🔄 MediaApiV2: Requesting TTS from voice-service: text="${text.substring(0, text.length > 50 ? 50 : text.length)}", voice="${voice ?? 'default'}", language="${language ?? 'not provided'}", instructions="${instructions != null ? instructions.substring(0, instructions.length > 50 ? 50 : instructions.length) : 'not provided'}"',
+        '🔄 MediaApiV2: Requesting TTS from voice-service: text="${text.substring(0, text.length > 50 ? 50 : text.length)}", voice="${voice ?? 'default'}", language="${language ?? 'not provided'}", instructions="${instructions != null ? instructions.substring(0, instructions.length > 50 ? 50 : instructions.length) : 'not provided'}", model="${model ?? 'not provided'}"',
       );
 
       // Call voice-service
@@ -529,6 +530,7 @@ class MediaApiV2 {
           'format': 'mp3',
           if (language != null) 'language': language,
           if (instructions != null) 'instructions': instructions,
+          if (model != null) 'model': model,
         },
         options: dio.Options(
           responseType: dio.ResponseType.bytes,