OpenCode Interrupt Plugin

Streaming TTS + voice interruption for OpenCode. Speaks responses as they arrive and detects when you talk over it. Walkie-talkie mode lets you type /ptt to redirect the model on the fly.

Installation

Add the plugin to your ~/.config/opencode/opencode.json:

{
  "plugin": [
    "opencode-interrupt-plugin"
  ]
}

The plugin handles TTS streaming, voice overlap detection, interrupt injection, and the /ptt walkie-talkie slash command.

Usage

Walkie-Talkie Mode

Type /ptt to toggle recording, speak, then type /ptt again to send:

The model is generating a response — TTS is playing
Type /ptt — generation aborts, TTS stops, mic starts recording
Speak your correction
Type /ptt again — audio is transcribed via Whisper and sent as your next message
The model responds to your correction

[Model speaking]  →  type /ptt → speak correction → type /ptt → model redirects

Requires whisper.cpp with the base model installed. Run the one-command install script:

bash scripts/install-whisper.sh

Or set OPENAI_API_KEY in your environment as a fallback.

Voice Interruption

When TTS is playing, just speak — the plugin detects your voice, stops TTS, and marks the session for correction injection.

Configuration

Setting	Type	Default	Description
`licenseKey`	string	none	License key (omit for free mode)
`micThreshold`	number	0.008	Voice activation threshold (free mode locked)
`timingWindowMs`	number	3000	How long after interrupt to accept corrections

Environment Variables

OPENAI_API_KEY — optional fallback for Whisper transcription (used if whisper CLI is not found)
WHISPER_MODEL — path to whisper.cpp model file (default: ~/.local/bin/ggml-base.bin)
EDGE_TTS_VOICE — TTS voice (default: en-US-AriaNeural)

How It Works

Server plugin: Monitors mic via sox (continuous PCM pipe), detects voice during TTS playback, injects correction context into the next LLM request
/ptt command: command.execute.before hook intercepts /ptt, toggles recording on/off, transcribes via whisper.cpp, and sends the transcript via session.prompt()