voice-ai - Build low-latency voice AI agents
Build voice AI agents using Deepgram STT, Groq LLM, Cartesia TTS, and Twilio Media Streams.
Tags
Updated: 2026-09-28Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Stream speech transcription
- Detect speech activity
- Generate low-latency responses
- Stream expressive speech synthesis
- Handle Twilio media streams
- Support bilingual conversations
Inputs
- User audio
- Provider API keys
- Language settings
- Voice settings
- Emotion settings
- Twilio call metadata
Outputs
- Speech transcripts
- Generated audio responses
- Interim transcription events
- Utterance end events
- Speech activity events
- Bidirectional call audio
Requirements
- Deepgram API access
- Groq API access
- Cartesia API access
- Twilio account access
- Async Python runtime
- WebSocket-capable server