Wire cf_voice modules when circuitforge-core#34 lands #9

New issue

Open

opened 2026-04-06 22:25:20 -07:00 by pyr0ball · 0 comments

pyr0ball commented

2026-04-06 22:25:20 -07:00

Owner

Summary

The current server has ad hoc Whisper/Piper/Precise wiring. As circuitforge_core.voice matures, these should be replaced:

Current code	Replace with
`voice_server.py:214-234` — `load_whisper_model()`, `transcribe_audio()`	`cf_voice.io.STTBackend`
`voice_server.py:237-251` — `generate_tts()` (stub)	`cf_voice.io.TTSBackend` (Piper)
`voice_server.py:253-322` — Precise wake word wiring	`cf_voice.wakeword.WakeWordDetector`
`voice_server_enhanced.py:160-217` — `SpeakerIdentification`	candidate for `cf_voice.speaker.SpeakerIDBackend`

Key note

HomeAssistantClient and IntentParser stay in Minerva — they are product-specific, not cf-core candidates.

The global mutable state (global whisper_model, etc.) must be refactored to injected instances before cf_voice modules can be wired in cleanly. Depends on FastAPI migration.

References

cf-dev review finding: Priority 3
circuitforge-core#34: cf_voice module
cf-orch#20: voice context service endpoint

## Summary The current server has ad hoc Whisper/Piper/Precise wiring. As `circuitforge_core.voice` matures, these should be replaced: | Current code | Replace with | |---|---| | `voice_server.py:214-234` — `load_whisper_model()`, `transcribe_audio()` | `cf_voice.io.STTBackend` | | `voice_server.py:237-251` — `generate_tts()` (stub) | `cf_voice.io.TTSBackend` (Piper) | | `voice_server.py:253-322` — Precise wake word wiring | `cf_voice.wakeword.WakeWordDetector` | | `voice_server_enhanced.py:160-217` — `SpeakerIdentification` | candidate for `cf_voice.speaker.SpeakerIDBackend` | ## Key note `HomeAssistantClient` and `IntentParser` stay in Minerva — they are product-specific, not cf-core candidates. The global mutable state (`global whisper_model`, etc.) must be refactored to injected instances before cf_voice modules can be wired in cleanly. Depends on FastAPI migration. ## References - cf-dev review finding: Priority 3 - circuitforge-core#34: cf_voice module - cf-orch#20: voice context service endpoint