setup-voice

Install a local voice interface for Claude Code on macOS Apple Silicon.

2|Updated Feb 23, 2026

Search Tags:#claude code #session management #text-to-speech #macos #speech-to-text #voice interface #hands-free coding

Ready to run

npx skills add https://github.com/MDGrey33/memnyx --skill setup-voice

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: setup-voice
Source: https://github.com/MDGrey33/memnyx/tree/main/.claude/skills/setup-voice
Command: npx skills add https://github.com/MDGrey33/memnyx --skill setup-voice

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill removes the friction of typing by letting you interact with Claude Code through a fully local voice pipeline, making hands-free prompting and spoken replies practical for coding sessions.

Core Features & Use Cases

Local speech-to-text: Records microphone input and transcribes it with mlx-whisper for fast, offline dictation.
Offline text-to-speech: Speaks responses with Kokoro neural voices, with separate personal and technical voice contexts.
Session-safe interaction: Supports single-turn and loop modes, mid-session context switching, and serialized playback so multiple sessions do not talk over each other.
Use cases: Dictate prompts while coding, run a continuous voice assistant loop, or switch to a clearer technical voice during development work.

Quick Start

Ask me to install and configure the setup-voice skill on my macOS Apple Silicon machine so I can talk to Claude Code aloud.

Frequently Asked Questions about setup-voice

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I set up a local voice interface for Claude Code on macOS?▼

You can achieve hands-free coding by setting up a local voice interface that records microphone input and transcribes it with mlx-whisper for offline dictation. The pipeline also uses Kokoro neural voices to speak responses back during your coding sessions.

Can I use mlx-whisper and Kokoro TTS for offline speech-to-text and text-to-speech with Claude Code?▼

Yes, mlx-whisper provides fast offline speech-to-text transcription, and Kokoro TTS models generate spoken replies using neural voices. This combination allows you to dictate prompts and hear spoken responses locally without relying on cloud services.

What dependencies do I need to install for a hands-free voice coding session on macOS Apple Silicon?▼

Required dependencies include Claude Code CLI, Homebrew, uv, Python 3.12, sox, and ffmpeg. These tools facilitate microphone capture, audio processing, and the execution of shell scripts needed for transcription and offline synthesis on macOS Apple Silicon.

Does the local voice pipeline support continuous voice loops and mid-session context switching?▼

The voice interface supports continuous voice assistant loops and mid-session context switching for dynamic coding workflows. It also includes serialized multi-session audio playback to ensure multiple concurrent sessions do not overlap or talk over each other.

How does the voice interface handle multiple concurrent Claude Code sessions?▼

Multiple concurrent sessions are managed through serialized multi-session audio playback, which prevents sessions from talking over each other. This ensures clear audio output when running several Claude Code voice interactions simultaneously.

Why use a local voice AI pipeline instead of cloud-based speech-to-text for coding prompts?▼

Using a local voice AI pipeline with mlx-whisper and Kokoro TTS keeps all speech-to-text and text-to-speech processing offline, ensuring data privacy and low latency. It also offers separate personal and technical voice contexts tailored for development work.