HiveMind Media Player

Turn any device into a remotely controlled OVOS media player via HiveMind.

hivemind-media-player ships HiveMindPlayerProtocol, a HiveMind agent protocol plugin (hivemind.agent.protocol) that runs the OVOS audio stack (ovos-audio and optional ovos-PHAL) locally and exposes it to any HiveMind client. Remote controllers send standard OCP (Open Voice OS Common Play) messages over the HiveMind encrypted WebSocket. The player device handles playback locally.

This is for devices that are not running a full OVOS instance. Think of a Raspberry Pi dedicated to being a networked speaker.

Architecture

remote controller           HiveMind (encrypted WebSocket)    this device
(Home Assistant,       <--------------------------------->    hivemind-core
 hivemind-player-ctl,                                         + HiveMindPlayerProtocol
 any OCP client)                                              + ovos-audio (TTS, OCP, VLC/MPV...)

The player agent answers no natural-language questions (natural_language_query yields only the end-of-query sentinel). Its sole role is to receive OCP/audio bus messages forwarded by hivemind-core and play them through the local audio stack.

Install

pip install hivemind-player-protocol

Optional extras (VLC, MPV backends):

pip install "hivemind-player-protocol[extras]"

Or from source:

git clone https://github.com/JarbasHiveMind/hivemind-media-player
cd hivemind-media-player
pip install -e .

Dependencies

The whole stack rides the ovos-bus-client 2.x line (HiveMind core 4.6.x requires it). That line, and the OVOS components updated to ride it (ovos-audio, OCP, ovos-plugin-manager, ovos-workshop), are currently published as prereleases.

pyproject.toml pins each dependency to its prerelease floor (for example ovos-bus-client>=2.0.0a3). A plain pip install then resolves the right versions. No --pre flag is needed. Do not add --pre. The floor pins opt into the prereleases, package by package, and keep resolution deterministic.

hivemind-core itself (AGPL) is not a runtime dependency. The plugin only imports its AgentProtocol base. The host process supplies hivemind-core. The test suite pulls it in through the [e2e] extra.

Quickstart

1. Configure hivemind-core

Edit ~/.config/hivemind-core/server.json on the player device:

{
  "agent_protocol": {
    "module": "hivemind-player-agent-plugin",
    "hivemind-player-agent-plugin": {}
  }
}

2. Configure ovos-audio

Edit ~/.config/mycroft/mycroft.conf on the same device:

{
  "play_wav_cmdline": "paplay %1",
  "play_mp3_cmdline": "mpg123 %1",
  "play_ogg_cmdline": "ogg123 -q %1",
  "tts": {
    "module": "ovos-tts-plugin-server"
  },
  "Audio": {
    "backends": {
      "OCP": {
        "type": "ovos_common_play",
        "preferred_audio_services": ["mpv", "vlc"],
        "active": true
      },
      "vlc": { "type": "vlc", "active": true },
      "mpv": { "type": "mpv",  "active": true }
    }
  }
}

3. Create a client credential

On the player device (where hivemind-core will run):

hivemind-core add-client
# note the Access Key and Password printed

4. Grant OCP permissions

# replace 3 with your Node ID from add-client
hivemind-core allow-msg "ovos.common_play.play" 3
hivemind-core allow-msg "ovos.common_play.pause" 3
hivemind-core allow-msg "ovos.common_play.resume" 3
hivemind-core allow-msg "ovos.common_play.stop" 3
hivemind-core allow-msg "ovos.common_play.next" 3
hivemind-core allow-msg "ovos.common_play.previous" 3
hivemind-core allow-msg "speak" 3

See Permissions for the full list.

5. Set the identity on the controller

On the device that will send commands:

hivemind-client set-identity \
  --key <access_key> --password <password> \
  --host <player_device_ip> --port 5678 --siteid player

6. Start hivemind-core on the player device

hivemind-core listen

7. Control playback

python hivemind-player-ctl.py play "http://example.com/audio/track.mp3"
python hivemind-player-ctl.py pause
python hivemind-player-ctl.py resume
python hivemind-player-ctl.py next

Home Assistant / Music Assistant Integration

With hivemind-homeassistant, HiveMind player devices appear as media players in Home Assistant. Music Assistant can then browse and play music to them.

Related projects:

• hivemind-homeassistant
• ovos-skill-music-assistant
• ovos-media-plugin-mass

hivemind-player-ctl

hivemind-player-ctl.py is a CLI to control a running player. It requires only hivemind_bus_client and click.

Usage: python hivemind-player-ctl.py [OPTIONS] COMMAND
Options:
  --key TEXT       Access key (or read from identity file)
  --password TEXT  Password (or read from identity file)
Commands:
  play URI          Start playback of a URI
  pause             Pause
  resume            Resume
  stop              Stop
  next              Next track
  prev              Previous track
  shuffle.set       Enable shuffle
  shuffle.unset     Disable shuffle
  repeat.set        Enable repeat-all
  repeat.one        Enable repeat-one
  repeat.unset      Disable repeat
  interactive       Interactive shell

Permissions

Core audio

Message	Purpose
speak	TTS output
mycroft.audio.is_alive	Health check
mycroft.audio.is_ready	Readiness check
mycroft.stop	Stop all audio

OCP (Open Voice OS Common Play)

Message	Purpose
ovos.common_play.play	Start playback
ovos.common_play.pause	Pause
ovos.common_play.resume	Resume
ovos.common_play.stop	Stop

Message	Purpose
ovos.common_play.next	Next track
ovos.common_play.previous	Previous track
ovos.common_play.player.status	Query player status
ovos.common_play.track_info	Query track info

Message	Purpose
ovos.common_play.playlist.queue	Queue a track
ovos.common_play.playlist.clear	Clear the queue
ovos.common_play.set_track_position	Seek

Message	Purpose
ovos.common_play.shuffle.set	Enable shuffle
ovos.common_play.shuffle.unset	Disable shuffle

Message	Purpose
ovos.common_play.repeat.set	Enable repeat
ovos.common_play.repeat.unset	Disable repeat
ovos.common_play.repeat.one	Repeat one

PHAL (optional)

Message	Purpose
mycroft.phal.is_alive	PHAL health
mycroft.phal.is_ready	PHAL readiness

Message	Purpose
mycroft.volume.get	Query volume
mycroft.volume.set	Set volume

Message	Purpose
mycroft.volume.increase	Volume up
mycroft.volume.decrease	Volume down
mycroft.volume.mute	Mute
mycroft.volume.unmute	Unmute

Running the tests

The suite is end-to-end. It stands up a real hivemind-core master in-process and drives the real player plugin over a real HiveMessageBusClient (through hivescope). It asserts that remote play, pause, and stop control commands round-trip from a satellite, through the deny-by-default ACL, to the player.

The audio playback backend is mocked. OCP and the legacy audio service are disabled, so no real audio plugin loads and nothing touches an audio device or the network.

pip install -e ".[test]"   # pulls hivescope + in-process hivemind-core ([e2e])
pytest tests/

The heavyweight e2e hosts live in the [e2e] extra. [test] includes them plus the test runner. hivescope and hivemind-core are required test dependencies. The suite never importorskips them.

Documentation

• docs/architecture.md: how the player protocol fits into HiveMind.
• docs/configuration.md: configuration reference.
• docs/permissions.md: full permissions reference.

License

Apache 2.0.