Voz
voicev0.3.1Inline voice mode in the chat: dictation and hands-free conversation with a live voice spectrum and 3 selectable engines
Description
Talk to the chat without opening any modal: the composer turns in place into a voice surface with a reactive spectrum, while the conversation thread stays visible. Two experiences share the same microphone: momentary dictation (speech to text, reviewed before sending) and interactive hands-free voice mode (you speak, it sends automatically, and the AI replies out loud). Three interchangeable engines are configurable from settings: 'cascaded' (default; uses the chat's native flow, so it KEEPS all tools, artifacts and KB and respects the user's Auto Tools state), 'audio-native' (speech-to-speech in a single call with gpt-audio, lower latency, pure conversation without tools) and 'elevenlabs-realtime' (low-latency duplex with interruption/barge-in via Conversational AI). Supports BYOK: each user can use their own ElevenLabs key (encrypted at rest) or the system's key.
Features (6)
Tap the mic: speak and the text appears in the input for review before sending
Press and hold: hands-free conversation with silence-based VAD; the AI replies out loud
Reactive voice visualizer in the composer itself (Web Audio), no modals. Different color per turn
The default engine inherits the Auto Tools state: artifacts, KB and tools remain available
cascaded · audio-native (gpt-audio, low latency) · elevenlabs-realtime (duplex with barge-in)
Each user can use their own ElevenLabs API key, encrypted with AES-256-GCM
Changelog
- Dictation and voice notes (#252): AIML transcription now uses «nova-2» by default; «whisper-large» kept processing without finishing and dictation was lost after 3 minutes. It can be changed with AIML_STT_MODEL.