Skip to content

README.md

Latest commit

 

History

History
197 lines (147 loc) · 5.57 KB

README.md

File metadata and controls

197 lines (147 loc) · 5.57 KB

Voice Mode Integration: VS Code

🔗 Official Documentation: VS Code Documentation
📦 Download/Install: Get VS Code
🏷️ Version Requirements: VS Code v1.99+ (with MCP preview feature)

Overview

Visual Studio Code is a popular open-source code editor from Microsoft. With the MCP preview feature in v1.99+, Voice Mode brings natural voice conversations to VS Code, enabling hands-free coding and verbal interactions with AI assistants.

Prerequisites

  • VS Code v1.99+ installed
  • Python 3.10 or higher
  • uv package manager (curl -LsSf https://astral.sh/uv/install.sh | sh)
  • OpenAI API key (or compatible service)
  • System audio dependencies installed (see main README)

Quick Start

# Enable MCP in VS Code settings
# Configure Voice Mode in .vscode/mcp.json
# Use Copilot Chat with voice commands

Installation Steps

1. Install VS Code

Download and install from: https://code.visualstudio.com/download

  • macOS: Download the .zip file and move to Applications
  • Linux: Use package manager or download .deb/.rpm package
  • Windows: Download the installer and run

2. Enable MCP Preview Feature

  1. Open VS Code Settings (Cmd+, or Ctrl+,)
  2. Search for "chat.mcp.enabled"
  3. Enable the checkbox for "Chat: Mcp: Enabled"
  4. Restart VS Code

3. Configure VS Code for Voice Mode

Configuration File Location: Create a .vscode/mcp.json file in your workspace root or user settings:

Add Voice Mode to MCP servers:

{
  "servers": {
    "voice-mode": {
      "command": "uvx",
      "args": ["voice-mode"],
      "env": {
        "OPENAI_API_KEY": "your-openai-key"
      },
      "type": "stdio"
    }
  }
}

Note: Using uvx means Voice Mode will be downloaded and run on-demand. No separate installation required!

4. Restart VS Code

After saving the configuration and enabling MCP, restart VS Code for changes to take effect.

5. Environment Variables (Optional)

For advanced configuration, you can set these environment variables:

# Required
export OPENAI_API_KEY="your-key"

# Optional - Custom STT/TTS endpoints (comma-separated lists)
export VOICEMODE_TTS_BASE_URLS="http://127.0.0.1:8880/v1,https://api.openai.com/v1"
export VOICEMODE_STT_BASE_URLS="http://127.0.0.1:2022/v1,https://api.openai.com/v1"

# Optional - Voice preferences (comma-separated lists)
export VOICEMODE_TTS_VOICES="af_sky,nova,alloy"
export VOICEMODE_TTS_MODELS="gpt-4o-mini-tts,tts-1-hd,tts-1"

Verification

  1. Check MCP Server Status:

    • Open Command Palette (Cmd+Shift+P or Ctrl+Shift+P)
    • Search for "Developer: Show Running Extensions"
    • Verify MCP servers are loaded

  2. Test Voice Mode:

    • Open VS Code
    • Open Copilot Chat (if available)
    • Type: "@voice-mode talk to me"
    • Try saying: "Hello, can you hear me?"

Usage Examples

Basic Voice Conversation

In Copilot Chat:
@voice-mode "Let's have a voice conversation"
You: "What does this code do?"
AI: [Speaks the explanation]

Voice-Enabled Coding

@voice-mode "Help me refactor this function using voice"
You: "Make this function async"
AI: [Explains changes verbally while showing code]

Troubleshooting

Voice Mode Not Available

  • Ensure MCP is enabled in VS Code settings (chat.mcp.enabled)
  • Check that .vscode/mcp.json exists and is valid JSON
  • Verify your OPENAI_API_KEY is set correctly

No Audio Input/Output

  • Check system audio permissions for VS Code
  • Run audio diagnostics: python scripts/diagnose-audio.py
  • Ensure microphone is not muted

MCP Feature Not Available

  • Ensure you have VS Code v1.99 or higher
  • Check that the MCP preview feature is enabled
  • Try using VS Code Insiders if stable version doesn't have it yet

Platform-Specific Notes

macOS

  • Grant microphone permissions to VS Code when prompted
  • Code Helper processes may also need microphone access

Linux

  • Install PulseAudio or PipeWire for audio support
  • May need: sudo apt-get install portaudio19-dev (Debian/Ubuntu)

Windows

  • Native Windows support requires WSL2
  • For best results, run VS Code from WSL2

Advanced Configuration

Using Local STT/TTS Services

To use local services for privacy:

  1. Start Kokoro TTS:

    # In VS Code terminal
uvx voice-mode kokoro-start

  2. Configure endpoints in .vscode/mcp.json:

    {
  "mcpServers": {
    "voice-mode": {
      "command": "uvx",
      "args": ["voice-mode"],
      "env": {
        "OPENAI_API_KEY": "your-key",
        "VOICEMODE_TTS_BASE_URLS": "http://127.0.0.1:8880/v1,https://api.openai.com/v1",
        "VOICEMODE_STT_BASE_URLS": "http://127.0.0.1:2022/v1,https://api.openai.com/v1"
      }
    }
  }
}

LiveKit Integration

For room-based voice conversations:

  • Configure LiveKit server in environment variables
  • Useful for collaborative coding sessions with voice

See Also

Need Help? Join our Discord community or check the FAQ