Kai Realtime Voice

Real-time voice streaming using MiniMax WebSocket API.

Setup

API key already configured in OpenClaw config. Uses same key as kai-minimax-tts.

Usage

Test WebSocket connection:

bash {baseDir}/scripts/kai_realtime.sh --test

Stream text to audio:

bash {baseDir}/scripts/kai_realtime.sh --stream "Hello world"

Requirements

• Python3 with websockets library
• WebSocket support
• MiniMax API key (auto-injected)

Output

Audio streamed to stdout or saved to file.