agent-media
davidj4tech/agent-media
Multi-channel mpv control surface for Termux: MCP server + HTTP/JSON API + mobile-first web UI with PWA + keyboard shortcuts
Instalação
npx convez-ai davidj4tech/agent-media1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
README
A tradução para o português deste README ainda está na fila.
agent-media
The agent's audio/music stack — gives coding agents a voice, whole-house playback control, and music awareness. Runs on any Linux box (x86 or ARM) or Termux on Android.
┌──────────────────────────────────────────────────────┐
│ packages/ │
│ ├── core/ intake → route → render → sink │ Python — the spine
│ ├── snapcast-room/ am-snap: snapcast routing CLI │ Python — plumbing
│ ├── voice-bridge/ STT (mic → text → intake) │ Python — sibling
│ └── astrotunes/ what to play │ Python — recommender
└──────────────────────────────────────────────────────┘
Packages
core/ — agent-media-core
The spine. Five subdirs that mirror the data flow:
intake/— event sources: Claude Code Stop/Notification hooks, Codex hook, Home Assistant SSE, Matrix, CLI (media say)route/— coordinator: content-type-aware Mopidy ducking/pause, local MPRIS browser pause, remote MPRIS via SSHrender/— TTS engines (edge / openai / qwen / realtime) with automatic fallbacksinks/— speech (long-running mpv broker over IPC) and music (Mopidy/MPD)state/— SQLite: now-playing, history, errors
Also exposes an MCP server (media-mcp) for tool-based control from Claude.
See packages/core/README.md for full
configuration reference (env vars, services, MPRIS, Snapcast).
snapcast-room/
The snapcast/pipewire plumbing that survived the restructure: am-snap, a
terse CLI over Snapcast's JSON-RPC for whole-house routing (join a room to a
channel, set volume, mute) across multiple snapservers. The former
agent-audio-relay's render engines, agent hooks, watcher and clip-server all
moved into core/; this is what remained. (aar-snap is kept as an alias.)
voice-bridge/
STT companion — mic capture → transcribe → submit_event into core's intake
pipeline.
astrotunes/ (in progress)
Given current planetary transits, time of day, mood, activity, and local weather, picks tracks and queues them via core's music sink.
Quick start
# 1. Clone and create a venv
git clone https://github.com/davidj4tech/agent-media
cd agent-media
python3 -m venv .venv && source .venv/bin/activate
# 2. Install core (editable — changes take effect immediately)
pip install -e packages/core
# 3. Configure — create ~/.config/agent-audio-relay.env
# See packages/core/README.md for all options. Minimal example:
cat > ~/.config/agent-audio-relay.env << 'EOF'
MEDIA_RENDER_ENGINE=edge
MEDIA_RENDER_VOICE=en-GB-SoniaNeural
EOF
# 4. Wire services and Claude Code hooks
media-setup
# 5. Source the tmux control surface (add to tmux.conf.local)
# source-file ~/.local/share/agent-media/media.tmux
Multi-host setup (mel → sp4r example)
mel is headless; sp4r is the laptop with speakers and a browser.
- mel renders TTS and routes audio to sp4r's Snapcast (
MEDIA_SPEECH_DEFAULT_TARGET=rooms) - sp4r runs
snapserver+snapclient+ Mopidy feedingam-music - When mel speaks, sp4r's Chrome/browser pauses automatically via SSH MPRIS:
set
MEDIA_MPRIS_SSH_HOSTS=sp4rin mel's env file
See packages/core/README.md → Remote MPRIS for details.
History
Assembled in May 2026 from previously separate repos:
davidj4tech/mpv-mcp→media-mcp(Node) → retired in favor ofcore.mcp_server(Python)davidj4tech/agent-audio-relay→packages/audio-relay/→ shrunk + renamed topackages/snapcast-room/(the rest absorbed intocore/)davidj4tech/tmux-voice-bridge→packages/voice-bridge/astrotunesnew