v0.8.1 · Full-Duplex Live Sync · 6 Agents · Mac & Windows Host

Hands-free voice
for your coding agent.

Talk to Claude Code, Codex, Antigravity, Ollama, Aider, or Google Gemini CLI directly from your phone. Multi-client live WebSocket sync keeps turns, tool approvals, and git diffs in real-time sync across your devices.

Download Mac DMG Download Windows Quickstart View Source
VoiceBridge desktop host control panel showing bridge health, Tailscale diagnostics, and pairing controls
Powerful Features

A seamless bridge between you and your code.

A tiny Node bridge drives CLI agents on your machine and streams everything to your phone, with browser voice or local Whisper streaming STT.

🎙️

Type or Speak

Use the chat interface to type, or tap the mic to speak. The agent's replies are formatted with rich markdown and read aloud sentence-by-sentence.

🤖

6 Coding Agents

Switch seamlessly between Claude Code, Codex, Antigravity, Ollama (fully local), Aider, and Gemini CLI. Each has configurable autonomy modes.

🔄

Full-Duplex Live Sync

Real-time WebSocket hub (/ws) syncs turns, streaming deltas, and tool approvals across multiple devices simultaneously.

⚡

Interactive Tool Approvals

Review and Approve or Reject shell commands and tool executions directly from interactive cards in chat and Talking Mode.

🔍

Git Diff Viewer Drawer

Inspect modified workspace files and review syntax-highlighted git diffs right from your phone's sliding drawer.

📞

Talking Mode & VAD

Hands-free voice loop with WebAudio Voice Activity Detection (VAD) for dynamic silence detection and auto-commit.

🎧

MediaSession Controls

Control conversations from your lock screen, Apple Watch, or Bluetooth headset with background audio keepalive.

📄

Transcript Export

Export any session transcript to Markdown (.md) or JSON (.json) via the native share sheet or instant file download.

🔒

Private by Default

Everything runs over your private Tailscale network via HTTPS. Your codebase never touches a third-party voice cloud.

🖥️

Desktop Host App

Install the Mac DMG or Windows preview build, choose one workspace, verify Tailscale, and pair your phone with a QR scan.

Architecture

Phone in, agent out.

Your phone handles the voice layer, while the heavy lifting (the CLI) stays on your Mac or PC.

[ iPhone Safari ] [ Your Mac or PC ] mic ─Web Speech or Whisper STT─▶ text ──https/Tailscale──▶ voicebridge ──spawn──▶ claude / codex / agy speaker ◀─speechSynthesis─ reply ◀────────────── reply ◀─────────── (agent CLI)
Get Started

Use the desktop app, or run from source.

The desktop app is the easiest host setup. The terminal flow stays available for developers who want full control.

Recommended

Desktop host app

Install the Mac DMG or Windows preview installer, select your project folder and agent, copy the Tailscale Serve command, verify the public URL, then scan the pairing QR from the mobile app.

Download voicebridge-0.2.1-arm64.dmg Download voicebridge-0.2.1-x64-setup.exe Checksums Mac setup guide Windows setup guide

Setup steps

1. Install an agent CLI and sign in on the host computer.
2. Open VoiceBridge, choose the project folder, and start the bridge.
3. Run the copied Tailscale Serve command and verify the public URL/token.
4. Scan the QR from the iOS/Android app.
git clone https://github.com/berkayturanci/voicebridge voicebridge
cd voicebridge && npm install

# Optional: Set up your environment (PROJECT_DIR, ACCESS_TOKEN)
cp .env.example .env

# Start the server (prints a QR code)
npm start

# In a new terminal tab: Expose securely to your phone
tailscale serve --bg http://127.0.0.1:8787

Tip: the desktop Network panel now distinguishes missing Tailscale, logged-out state, DNS/network failure, HTTP status, timeout, and token mismatch.

FAQ

Common Questions

Is the voice feature actually free?
Yes. Browser speech-to-text and text-to-speech need no paid voice provider, and Whisper modes can keep transcription local on your own machine. You only pay for actual cloud agent usage where that agent CLI requires it.
Does my code leave my machine?
VoiceBridge acts only as a local proxy. Your agent CLI runs on your machine, and traffic goes over your private Tailscale network. Claude, Codex, and Antigravity may still call their vendor clouds; Ollama and local Whisper can stay fully local.
Why do I need Tailscale and HTTPS?
Modern browsers (Safari, Chrome) strictly enforce security and completely disable microphone access on non-secure origins (like standard `http://192.168.x.x`). Tailscale Serve provides a real, trusted HTTPS certificate over your private network, instantly unlocking mic access.
Can I use Whisper instead?
Absolutely. Use `STT_MODE=whisper` with `STT_CMD` for batch transcription, or `STT_MODE=whisper-stream` with `STT_STREAM_URL` for a local WebSocket transcriber that works with hands-free talking mode.