feat: lots of speech stuff

This commit is contained in:
Brooklyn Nicholson
2026-05-01 19:28:02 -05:00
parent 9f3d393a4d
commit d5d7b5c6dc
41 changed files with 1405 additions and 361 deletions
+9 -2
View File
@@ -830,7 +830,7 @@ DEFAULT_CONFIG = {
"stt": {
"enabled": True,
"provider": "local", # "local" (free, faster-whisper) | "groq" | "openai" (Whisper API) | "mistral" (Voxtral Transcribe)
"provider": "local", # "local" (free, faster-whisper) | "groq" | "openai" (Whisper API) | "mistral" (Voxtral Transcribe) | "elevenlabs" (Scribe)
"local": {
"model": "base", # tiny, base, small, medium, large-v3
"language": "", # auto-detect by default; set to "en", "es", "fr", etc. to force
@@ -841,6 +841,12 @@ DEFAULT_CONFIG = {
"mistral": {
"model": "voxtral-mini-latest", # voxtral-mini-latest, voxtral-mini-2602
},
"elevenlabs": {
"model_id": "scribe_v2", # scribe_v2, scribe_v1
"language_code": "", # auto-detect by default; set to "eng", "spa", "fra", etc. to force
"tag_audio_events": False,
"diarize": False,
},
},
"voice": {
@@ -1791,9 +1797,10 @@ OPTIONAL_ENV_VARS = {
"category": "tool",
},
"ELEVENLABS_API_KEY": {
"description": "ElevenLabs API key for premium text-to-speech voices",
"description": "ElevenLabs API key for premium text-to-speech voices and Scribe transcription",
"prompt": "ElevenLabs API key",
"url": "https://elevenlabs.io/",
"tools": ["elevenlabs_tts", "voice_transcription"],
"password": True,
"category": "tool",
},
+6 -1
View File
@@ -280,7 +280,12 @@ _SCHEMA_OVERRIDES: Dict[str, Dict[str, Any]] = {
"stt.provider": {
"type": "select",
"description": "Speech-to-text provider",
"options": ["local", "openai", "mistral"],
"options": ["local", "groq", "openai", "mistral", "xai", "elevenlabs"],
},
"stt.elevenlabs.model_id": {
"type": "select",
"description": "ElevenLabs Scribe model",
"options": ["scribe_v2", "scribe_v1"],
},
"display.skin": {
"type": "select",