English ｜ 简体中文
Phone → PC · Wireless Microphone
Turn your mobile device into a Windows PC microphone.
Chromium-based browser only. Zero install on device. No cloud.
How It Works · Features · Quick Start · Architecture · Dev · Status · FAQ
┌─────────────────────────────────────────────────┐
│ 📱 MOBILE DEVICE (Chromium) │
│ │
│ Open URL from terminal │
│ Tap "Start Microphone" │
│ Start speaking │
│ │
│ getUserMedia │
│ → MediaStreamTrackProcessor │
│ → AudioEncoder (opus · 48kHz · mono) │
│ → WebSocket send │
│ │ │
└───────────────────────│─────────────────────────┘
│ HTTPS + WebSocket
│ (local network · ~80ms)
│
┌───────────────────────│─────────────────────────┐
│ ▼ │
│ 🖥️ WINDOWS PC (darkmic.exe) │
│ │
│ WebSocket receive │
│ → opusscript decode (opus → PCM) │
│ → WinMM waveOut → VB-CABLE Input │
│ → VB-CABLE Output (system virtual mic) │
│ │
│ ✅ Any app sees it as a normal microphone │
└─────────────────────────────────────────────────┘
Audio flows one-way (phone → PC). End-to-end latency ~ 80ms.
| 📱 Zero install on device | Just a Chromium browser — no app store needed |
| 🔒 Local network only | No cloud, no data leaves your home |
| ⚡ ~80ms latency | WebCodecs AudioEncoder, measured via ping/pong |
| 🎚️ Auto Gain Control | DynamicsCompressorNode smooths speaking volume |
| 🎵 Adaptive bitrate | Auto-switches 32↔64kbps based on latency |
| 📊 Live latency | Real-time one-way ms on dashboard + phone |
| 🔊 Opus codec | Speech-optimized, efficient bandwidth |
| 🔌 VB-Cable output | System-level virtual microphone device |
| 📦 Single .exe | Packaged via pkg, FFmpeg bundled alongside |
| 🖥️ Windows 10/11 | Any mobile device with Chromium browser |
| 🌐 QR connect | Scan QR on PC dashboard to connect phone |
Prerequisites
# Install system dependencies (Windows)
winget install ffmpeg
winget install OpenSSL.LightImportant
Also required: VB-Cable (virtual audio driver) and a Chromium-based browser (Chrome, Edge, Brave, etc.) on both PC and mobile. Both devices on the same WiFi.
Run
git clone https://github.com/darkLordIceCream/darkmic.git
cd darkmic
pnpm install
pnpm run devThe terminal shows your LAN IPs:
➜ Dashboard: https://localhost:3000
➜ Phone: https://192.168.1.100:3000/phone
Connect
| ① | Open https://localhost:3000 on your PC — the dashboard appears with a QR code |
| ② | Scan the QR code with your mobile device's browser, or open the Phone URL manually |
| ③ | Tap Advanced → Proceed (self-signed cert — expected) |
| ④ | Tap "Start Microphone" → allow mic permission |
| ⑤ | Speak — audio streams to your PC's VB-Cable virtual microphone |
Tech Stack
| Layer | Technology |
|---|---|
| Server runtime | Node.js + TypeScript |
| HTTP | Express |
| WebSocket | ws |
| Audio encode (browser) | WebCodecs AudioEncoder (opus) |
| Audio decode (server) | opusscript (pure JS) |
| Audio output | WinMM waveOut via koffi |
| Packaging | @yao-pkg/pkg |
Project Layout
darkmic/
├── src/ # PC server (TypeScript)
│ ├── index.ts # HTTPS + WebSocket + Express
│ ├── cert.ts # Self-signed SSL cert generation
│ ├── audio.ts # AudioPipe factory (3 modes)
│ └── wasapi.ts # WinMM waveOut → VB-Cable
├── public/ # Web app pages
│ ├── pc.html # PC dashboard (QR code, metrics, event log)
│ ├── index.html # Phone UI
│ └── client.js # WebCodecs + WebSocket + auto-reconnect
├── scripts/ # Dev utilities
├── AGENTS.md # AI agent instructions
├── feature_list.json # Feature tracker
├── progress.md # Session log
└── init.sh # Verification script
Future: If lower latency is needed, upgrade transport to WebRTC (
RTCPeerConnection). WebSocket becomes signaling only.
./init.sh # Full verification (install + typecheck + build + audio test)
pnpm run dev # Dev server with hot reload
pnpm run typecheck # TypeScript type check only
pnpm run build # Compile TypeScript to dist/
pnpm run start # Production server (node dist/index.js)
pnpm test:audio # Audio pipe test → file mode
pnpm test:audio -- ffplay # Audio pipe test → speaker playback
pnpm test:audio -- wasapi # Audio pipe test → VB-Cable output| ID | Feature | Completed | |
|---|---|---|---|
| F-001 | Project scaffold + HTTPS certs | 2026-05-13 | done |
| F-002 | WebCodecs + WebSocket pipeline | 2026-05-13 | done |
| F-003 | opusscript decode → WinMM → VB-Cable | 2026-05-16 | done |
| F-004 | QR code + PC dashboard + auto-reconnect | 2026-05-16 | done |
| F-005 | AGC + adaptive bitrate + latency measurement | 2026-05-16 | done |
| F-006 | WebRTC P2P transport | — | deferred |
| F-007 | Windows pkg packaging | — | todo |
| F-008 | Installer + system tray | — | todo |
Why does the browser show a privacy warning?
The server generates a self‑signed SSL certificate on first run. Chrome warns because it isn't signed by a public CA. Your connection is still encrypted — perfectly safe for local network use. Tap Advanced → Proceed to continue.
Can I use Safari or Firefox?
No. This project requires a Chromium-based browser (Chrome, Edge, Brave, Opera, etc.) on both mobile and PC. It uses WebCodecs
AudioEncoder which is only implemented in Chromium.
Can I use this over the internet?
No. darkmic is designed for local network only (same WiFi subnet). There is no STUN/TURN, no cloud relay, no NAT traversal.
English ｜ 简体中文
Built for when you need a mic but don't have one. © 2026 darkLordIceCream