Voice-to-text for vibe coding. Record your voice, auto-transcribe with OpenAI Whisper, and paste directly into any VS Code input — including AI chat boxes (Claude Code, Copilot Chat, ChatGPT, etc.)
- Press
PgDnto start recording — status bar turns red - Press
PgDnagain to stop — transcription happens automatically - Text is copied to clipboard and pasted at your cursor position
- Tap 🎤 Record in the VibeMic panel
- Speak — button turns red while recording
- Tap ⏹ Stop — transcription happens automatically
- Tap 📋 Copy to copy to clipboard, then paste wherever you need
- One-key recording —
PgDnto start and stop (desktop) - Auto-paste — transcribed text is automatically pasted at cursor position (desktop)
- Auto-copy — also copies to clipboard as backup (desktop)
- Mobile support — works in code-server (iPad / iPhone) with a webview copy button
- Multi-language — supports Cantonese, Mandarin, English, Japanese, and 97 other languages
- Cross-platform — works on Linux and macOS
- OpenAI API key with Whisper access ($0.006/min — very cheap)
- sox for audio recording:
- Linux:
sudo apt install sox - macOS:
brew install sox
- Linux:
- Install the extension
- Open VS Code Settings (
Ctrl+,) - Search
vibemic - Set your OpenAI API key
- (Optional) Set language code (e.g.
zhfor Chinese,enfor English, or leave empty for auto-detect)
- Install the
.vsixmanually into code-server - Open the VibeMic tab in the bottom panel (next to Terminal)
- Set your OpenAI API key in Settings
| Action | Key |
|---|---|
| Start recording | PgDn |
| Stop recording | PgDn |
AI chat boxes in VS Code don't support voice input. The built-in VS Code Speech extension only works with Copilot Chat. VibeMic fills the gap — it works with any input field in VS Code.
Built for the vibe coding era. Stop typing, start talking.
MIT