HiveMind Voice Satellite

HiveMind Voice Satellite is a full-stack OVOS voice satellite. Microphone capture, VAD, wakeword detection, STT, and TTS all run on this device. The satellite transcribes speech locally and sends only the resulting text utterance to the hive. Responses arrive as text and the satellite speaks them locally. This satellite needs the most compute of the satellite family. It puts the least load on the server and works fully offline once set up.

Satellite spectrum

Use voice-sat when the device has CPU/GPU to spare, bandwidth is limited, or audio privacy matters.

Install

pip install HiveMind-voice-sat
# Linux: ALSA or SoundDevice microphone support
pip install HiveMind-voice-sat[linux]
# macOS
pip install HiveMind-voice-sat[mac]

60-second quickstart

1. Add a client on the hive (run this on the server):

hivemind-core add-client --name my-voice-sat
# outputs: Access Key and Password, copy them

2. Run the satellite (on this device):

hivemind-voice-sat --host <hive-host> --key <access-key> --password <password>

Say your wakeword. The satellite transcribes locally and sends the utterance to the hive.

Minimal configuration

The satellite reads ~/.config/mycroft/mycroft.conf (standard OVOS config). Override STT, TTS, VAD, and wakeword plugins there:

{
  "listener": {
    "VAD": {
      "module": "ovos-vad-plugin-silero"
    },
    "wake_word": "hey_mycroft",
    "hey_mycroft": {
      "module": "ovos-ww-plugin-vosk"
    }
  },
  "stt": {
    "module": "ovos-stt-plugin-server",
    "ovos-stt-plugin-server": { "url": "https://stt.openvoiceos.org/stt" }
  },
  "tts": {
    "module": "ovos-tts-plugin-server",
    "ovos-tts-plugin-server": { "host": "https://tts.openvoiceos.org" }
  }
}

All plugin slots are swappable via ovos-plugin-manager. See docs/configuration.md for the full reference.

CLI reference

Usage: hivemind-voice-sat [OPTIONS]
  connect to HiveMind
Options:
  --host TEXT       hivemind host (ws:// or wss://)
  --key TEXT        Access Key
  --password TEXT   Password for key derivation
  --port INTEGER    HiveMind port number (default 5678)
  --selfsigned      accept self-signed certificates
  --siteid TEXT     location identifier for message.context
  --help            Show this message and exit.

Documentation

Full zero-to-hero docs live in docs/:

• Overview & architecture spectrum

• Getting started

• Configuration reference

• Architecture (advanced)

• Deployment (systemd / Raspberry Pi)

• Troubleshooting

• Testing & development

Development

git clone https://github.com/JarbasHiveMind/HiveMind-voice-sat
cd HiveMind-voice-sat
pip install -e ".[test]"   # lightweight smoke deps
pytest tests/test_smoke.py
pip install -e ".[e2e]"    # full hivemind-core + 2.x stack for end-to-end
pytest tests/e2e/

pyproject.toml is the single packaging source of truth. The dependency stack now runs on ovos-bus-client 2.x (see docs/architecture.md). Prerelease dependencies are pinned as minimum versions, so pip resolves them without --pre.

Related

License

Apache-2.0. See LICENSE.