Files
voice-assistant/docs/voice-assistant.mmd

29 lines
1016 B
Plaintext

flowchart LR
subgraph OnDevice["On-device"]
ESP32[ESP32-S3<br/>wake word · mic · LED]
end
subgraph Lab["Home lab (.13 · .30 · .35)"]
MQTT[(MQTT<br/>mosquitto :1883)]
WHISPER[Whisper STT<br/>faster-whisper]
AGENT[Jervis Agent<br/>FastAPI :8501]
HA[Home Assistant<br/>.30 :8123]
OMNI[OmniRoute<br/>LLM gateway :20129]
MOPIDY[Mopidy · Spotify<br/>music :6600]
APPRISE[Apprise<br/>notifications .35]
SNAP[Snapcast<br/>multi-room :1780]
PIPER[Piper TTS<br/>:10200]
SPEAKERS[House Speakers<br/>whole home]
end
ESP32 -->|audio stream| MQTT
MQTT -.->|speech frames| WHISPER
WHISPER -->|transcript| AGENT
AGENT -->|control · todo · timers| HA
AGENT -->|intent / Q&A| OMNI
AGENT -.->|play · pause · volume| MOPIDY
AGENT -.->|family messages| APPRISE
AGENT -->|reply text| SNAP
SNAP -.->|synthesize speech| PIPER
PIPER -->|speech wav| SNAP
SNAP -->|audio out| SPEAKERS