navig-audio

NAVIG Audio — the voice subsystem (text-to-speech, speech-to-text, wake-word, streaming pipeline). A first-party navig plugin (free, toggleable), extracted from core navig.voice. Phase 3 folds AI audio generation in here too.

What it does

• TTS / STT engines (edge-tts, faster-whisper, OpenAI) behind navig.voice.*.
• Powers the navig voice CLI, gateway Telegram voice notes, agent voice input, and inbox audio transcription.

How it wires (compat shims)

Core keeps thin navig.voice.* compatibility shims that re-export navig_audio.voice.*. So every existing consumer works unchanged when navig-audio is installed, and degrades gracefully (ImportError → feature off) when it isn't. The heavy deps (edge-tts / faster-whisper / openai) live here, not in core — a bare pip install navig stays lean.

Install

Bundled with a full navig install (the voice/audio extra); standalone: navig store install pip:navig-audio. Requires navig-core in the same environment.

Roadmap (Phase 3)

navig-audio joins navig-image · navig-video · navig-text as the media-type plugin family, each owning its own generate. Audio generation + narration (feeding navig-video and navig-social) land here, coordinated with siblings via core's refs library + event bus.