Module voice 

Voice pipeline components for ClawFT.

All types are gated behind the voice feature flag. Sub-features (voice-stt, voice-tts, voice-vad, voice-wake) control which pipeline components are compiled.

Re-exports

pub use config::VoicePipelineConfig;

pub use models::ModelDownloadManager;

pub use channel::VoiceChannel;

pub use channel::VoiceStatus;

pub use echo::EchoCanceller;

pub use echo::EchoCancellerConfig;

pub use events::VoiceWsEvent;

pub use noise::NoiseSuppressor;

pub use noise::NoiseSuppressorConfig;

pub use quality::AudioMetrics;

pub use quality::analyze_frame;

pub use talk_mode::TalkModeController;

pub use wake::WakeWordConfig;

pub use wake::WakeWordDetector;

pub use wake::WakeWordEvent;

pub use wake_daemon::WakeDaemon;

Modules

capture

Audio capture (microphone input) via cpal.

channel

Voice channel adapter implementing ChannelAdapter.

cloud_stt

Cloud-based speech-to-text providers.

cloud_tts

Cloud-based text-to-speech providers.

commands

Voice command shortcuts.

config

Voice-specific configuration helpers.

echo

Echo cancellation for voice pipeline.

events

WebSocket event types for voice status reporting.

fallback

STT/TTS fallback chains: local-first with cloud fallback.

models

Model download and cache management.

noise

Noise suppression for voice pipeline.

playback

Audio playback (speaker output) via cpal.

quality

Audio quality monitoring for voice pipeline.

stt

Speech-to-text via sherpa-rs streaming recognizer.

talk_mode

Talk Mode controller for continuous voice conversation.

transcript_log

Voice session transcript logging.

tts

Text-to-speech via sherpa-rs streaming synthesizer.

vad

Voice Activity Detection using Silero VAD.

wake

Wake word detection for “Hey Weft” trigger phrase.

wake_daemon

Background daemon for continuous wake word detection.