Skip to content
agentusehub

Debug code hands-free by voice with Claude Code

Author built a local speech-to-text and text-to-speech stack hooked to Claude Code via MCP, and used it to debug a WhatsApp integration hands-free while brushing teeth.

Time saved
Not reported
Setup time
Not reported
Running cost
Not reported
Cost range
Not reported

Use case overview

Author built a local speech-to-text and text-to-speech stack hooked to Claude Code via MCP, and used it to debug a WhatsApp integration hands-free while brushing teeth.

Reported experience

  1. What they did

    The author spent an afternoon building a local voice stack on an M4 Pro: Parakeet v3 TDT for speech-to-text and Pocket TTS for speech-to-speech output, joined to Claude Code through a custom MCP. They hacked AirPods stem clicks to be captured, so a click interrupts Claude and starts listening. Claude narrates its work in short, frequent messages, and the session is streamed by AirPlay to a TV.

  2. How it went

    The author had it debug a custom WhatsApp integration hands-free while brushing teeth, and calls it fun and easy on the back. The source gives no numbers or failure details.

  3. Worth knowing

    Getting the AirPods stem click captured took some hacking, and the stack ran on an M4 Pro, so expect to tinker with hardware input and local model speed.

Example prompts

Try with Claude Code
Help me set up hands-free voice control for Claude Code on my [Mac/Windows/Linux] machine, using local speech-to-text and text-to-speech connected through an MCP server. Walk me through the setup step by step. Finished means I can speak a request, hear a spoken reply, and use it to debug [project or integration name]. Show me any command that installs software before running it.
Software & DevOps

Port a Chrome extension to Safari

The author had ChatGPT take the Session Buddy extension source from Chrome, port it to Safari and install it, taking about 20 minutes.

ChatGPT