feat: add speech2text (#1855)

- A new stt driver in `AIInterfaceService` and `PuterAIModule`.
- added methods for audio transcription and translation in the speech-to-text interface.
- updated cost mapping for stt models in `openAiCostMap.ts`.
- Updated permissions and interfaces to support new speech-to-text features.

Author: jelveh

package-lock.json

@@ -61,6 +61,7 @@
     "multer": "^2.0.2",
     "multi-progress": "^4.0.0",
     "murmurhash": "^2.0.1",
+    "music-metadata": "^7.14.0",
     "nodemailer": "^6.9.3",
     "on-finished": "^2.4.1",
     "openai": "^6.7.0",

src/backend/package.json

@@ -25,6 +25,7 @@ const default_implicit_user_app_permissions = {
     'driver:puter-image-generation': {},
     'driver:puter-video-generation': {},
     'driver:puter-tts': {},
+    'driver:puter-speech2txt': {},
     'driver:puter-apps': {},
     'driver:puter-subdomains': {},
     'driver:temp-email': {},
@@ -60,6 +61,8 @@ const implicit_user_app_permissions = [
             'driver:puter-chat-completion:complete': {},
             'driver:puter-image-generation:generate': {},
             'driver:puter-video-generation:generate': {},
+            'driver:puter-speech2txt:transcribe': {},
+            'driver:puter-speech2txt:translate': {},
             'driver:puter-analytics:create_trace': {},
             'driver:puter-analytics:record': {},
         },

src/backend/src/data/hardcoded-permissions.js

@@ -205,6 +205,50 @@ class AIInterfaceService extends BaseService {
                 },
             }
         })
+
+        col_interfaces.set('puter-speech2txt', {
+            description: 'Speech to text transcription and translation.',
+            methods: {
+                list_models: {
+                    description: 'List available speech-to-text models.',
+                    result: { type: 'json' },
+                },
+                transcribe: {
+                    description: 'Transcribe audio into text.',
+                    parameters: {
+                        file: { type: 'file' },
+                        model: { type: 'string', optional: true },
+                        response_format: { type: 'string', optional: true },
+                        language: { type: 'string', optional: true },
+                        prompt: { type: 'string', optional: true },
+                        temperature: { type: 'number', optional: true },
+                        logprobs: { type: 'flag', optional: true },
+                        timestamp_granularities: { type: 'json', optional: true },
+                        stream: { type: 'flag', optional: true },
+                        chunking_strategy: { type: 'string', optional: true },
+                        known_speaker_names: { type: 'json', optional: true },
+                        known_speaker_references: { type: 'json', optional: true },
+                        extra_body: { type: 'json', optional: true },
+                    },
+                    result: { type: 'json' },
+                },
+                translate: {
+                    description: 'Translate audio into English text.',
+                    parameters: {
+                        file: { type: 'file' },
+                        model: { type: 'string', optional: true },
+                        response_format: { type: 'string', optional: true },
+                        prompt: { type: 'string', optional: true },
+                        temperature: { type: 'number', optional: true },
+                        logprobs: { type: 'flag', optional: true },
+                        timestamp_granularities: { type: 'json', optional: true },
+                        stream: { type: 'flag', optional: true },
+                        extra_body: { type: 'json', optional: true },
+                    },
+                    result: { type: 'json' },
+                },
+            },
+        });
     }
 }