Lightning Voice

Complete AI voice pipeline. Text-to-Speech, Speech-to-Text, and voice bridge integration for natural AI conversations.

v1.0.0 Production

Voice Pipeline

    ┌─────────────┐     ┌──────────────┐     ┌─────────────┐
    │   Microphone │────▶│     STT      │────▶│   Claude    │
    │   (Input)    │     │ (Whisper/VA) │     │   (Think)   │
    └─────────────┘     └──────────────┘     └──────┬──────┘
                                                     │
    ┌─────────────┐     ┌──────────────┐            │
    │   Speaker   │◀────│     TTS      │◀───────────┘
    │   (Output)  │     │ (Piper/Edge) │
    └─────────────┘     └──────────────┘
    
    Voice In → Transcribe → AI Process → Synthesize → Voice Out

Components

TTS Engine

Multiple TTS backends including Piper, Edge TTS, and ElevenLabs for natural speech synthesis.

STT Engine

Whisper-based speech recognition with real-time transcription and noise filtering.

Voice Bridge

Connects voice I/O to Claude via Lightning Conduit for seamless conversation flow.

Voice Server

WebSocket server for real-time bidirectional voice streaming.

Integration

Hands-Free AI: Lightning Voice enables natural voice conversations with Claude. Perfect for accessibility, mobile use, or working while your hands are busy.

# Start voice session
POST /api/voice/start
{
  "tts_engine": "piper",
  "stt_engine": "whisper",
  "voice": "en_US-amy-medium",
  "claude_project": "l_richard"
}

# Voice flows through the pipeline automatically
# Speak → Transcribe → Claude → Synthesize → Hear response