Polyphon

pending

by Corey Daley

Connect to a running Polyphon instance and have multi-voice AI conversations from within your vault.

Updated 5d agoDiscovered via Obsidian Unofficial Plugins
View on GitHub

Polyphon for Obsidian

CI GitHub release License Follow on X

An Obsidian plugin that connects to a running Polyphon instance, letting you have multi-voice AI conversations from within your vault.

Requirements

  • Polyphon must be installed and running locally (desktop app)
  • Obsidian desktop — mobile is not supported

Installation

The plugin is not yet listed in the Obsidian community plugin directory. Install it manually or via BRAT.

Manual installation (recommended)

  1. Download main.js, manifest.json, and styles.css from the latest release
  2. Create the folder <vault>/.obsidian/plugins/polyphon/ if it doesn't exist
  3. Copy the three files into that folder
  4. Enable the plugin in Settings → Community plugins

BRAT

Install BRAT, then add polyphon-ai/obsidian-polyphon as a beta plugin. BRAT will keep it up to date automatically.

Configuration

Open Settings → Polyphon after enabling the plugin:

SettingDefaultDescription
Host127.0.0.1Hostname or IP of your Polyphon instance
Port7432TCP port Polyphon is listening on
API token(auto)Use Read local token to load from the running instance, or paste manually
Persist conversationsOffSave and restore history across restarts
Debug modeOffLog raw JSON-RPC frames to the console

The API token is stored at ~/Library/Application Support/Polyphon/api.key on macOS. The Read local token button in settings loads it automatically.

Usage

  1. Start Polyphon
  2. Open the sidebar — click the message icon in the ribbon, or use Cmd/Ctrl+P → Polyphon: Open sidebar
  3. The plugin connects automatically; the status badge shows online when ready
  4. Select a composition from the dropdown
  5. Choose or create a session
  6. Type a message and press Send or Enter

Each voice in the composition responds in the unified thread, labeled by name. Use @VoiceName to target a specific voice.

Voice targeting

Type @ in the message field to see a dropdown of voices. Select one to direct your message to that voice only.

Session context

The plugin automatically includes your current open file path in the first message of each session, giving the AI voices context about what you're working on.

Development

npm install
npm run dev    # watch mode — auto-copies to vault/.obsidian/plugins/polyphon/
npm run build  # production build
npm test       # unit + integration tests
npm run lint   # ESLint

See CONTRIBUTING.md for details on the test suite and architecture.

Community

License

MIT — see LICENSE

For plugin developers

Search results and similarity scores are powered by semantic analysis of your plugin's README. If your plugin isn't appearing for searches you'd expect, try updating your README to clearly describe your plugin's purpose, features, and use cases.