Text-to-Speech (TTS)
此内容尚不支持你的语言。
Overview
Section titled “Overview”Clawdbot can respond with voice messages using text-to-speech synthesis.
Supported TTS Engines
Section titled “Supported TTS Engines”| Engine | Quality | Speed | Cost |
|---|---|---|---|
| OpenAI TTS | High | Fast | Paid |
| ElevenLabs | Very High | Medium | Paid |
| Google TTS | Medium | Fast | Free tier |
| Local (pyttsx3) | Low | Fast | Free |
Configuration
Section titled “Configuration”OpenAI TTS
Section titled “OpenAI TTS”TTS_ENABLED=trueTTS_ENGINE=openaiOPENAI_API_KEY=sk-xxxxxTTS_VOICE=alloy # alloy, echo, fable, onyx, nova, shimmerElevenLabs
Section titled “ElevenLabs”TTS_ENABLED=trueTTS_ENGINE=elevenlabsELEVENLABS_API_KEY=xxxxxTTS_VOICE=voice_idGoogle TTS
Section titled “Google TTS”TTS_ENABLED=trueTTS_ENGINE=googleTTS_LANGUAGE=enLocal (pyttsx3)
Section titled “Local (pyttsx3)”TTS_ENABLED=trueTTS_ENGINE=localChannel Support
Section titled “Channel Support”| Channel | Voice Support |
|---|---|
| Telegram | ✅ Voice messages |
| Discord | ✅ Voice channels |
| ⚠️ Limited | |
| Slack | ❌ |
When enabled, add --voice or configure auto-TTS:
# All responses as voiceTTS_AUTO=true
# Only when requestedTTS_AUTO=falseTTS_TRIGGER=voice # User says "voice" to get audio