LLM Gateway Connector

Connect GitHub Copilot Chat to your company's LLM gateway with zero configuration. Automatically discovers available models and proxies requests through litellm.

Features

• 🔄 Automatic Model Discovery - Fetches available models from your gateway
• 🚀 Zero Config - Just add your gateway URL and API key
• 🔌 Built-in Proxy - Runs litellm proxy in the background
• 🔒 Secure - API keys stored in VS Code's secure storage
• 🎯 Compatible - Works with any OpenAI-compatible LLM gateway
• 🔍 Smart Detection - Automatically finds litellm installation across different Python versions

Installation

From VS Code Marketplace

1. Open VS Code
2. Go to Extensions (Cmd/Ctrl+Shift+X)
3. Search for "LLM Gateway Connector"
4. Click Install

From VSIX file

code --install-extension llm-gateway-connector-0.1.0.vsix

Configuration

After installation:

1. Open Settings (Cmd/Ctrl + ,)

2. Search for "LLM Gateway"

3. Configure the following:

   • Gateway URL: Your company's LLM gateway endpoint (e.g., https://llmgateway.company.com)
   • API Key: Your authentication key
   • Proxy Port (optional): Local port for proxy (default: 8000)
   • Litellm Path (optional): Custom path to litellm executable (auto-detected if empty)

4. Reload VS Code (Cmd/Ctrl+Shift+P → "Developer: Reload Window")

The extension will automatically:

• Detect or install litellm
• Fetch available models from your gateway
• Start the proxy server
• Register models with GitHub Copilot Chat

Requirements

• GitHub Copilot subscription
• Python 3.8+ (for litellm)
• Network access to your LLM gateway

Extension Settings

• llmGateway.url: Your LLM Gateway base URL
• llmGateway.apiKey: API key for authentication
• llmGateway.proxyPort: Local port for the proxy (default: 8000)
• llmGateway.autoRefreshModels: Auto-fetch models on startup
• llmGateway.dropParams: Parameters to drop for compatibility

Commands

• LLM Gateway: Refresh Models - Fetch latest models from gateway
• LLM Gateway: Restart Proxy - Restart the litellm proxy
• LLM Gateway: Show Status - Show extension logs

Usage

After configuration, your gateway models will appear in the Copilot Chat model selector. Select any model to start chatting!

Troubleshooting

Models not appearing?

1. Check the Output panel (LLM Gateway channel)
2. Verify your gateway URL and API key
3. Run "LLM Gateway: Refresh Models"

Proxy not starting?

1. Ensure Python and pip are installed
2. Run "LLM Gateway: Restart Proxy"
3. Check the Output panel for errors

Privacy

This extension:

• Only communicates with your configured gateway
• Stores API keys securely in VS Code
• Runs the proxy locally on your machine
• Does not send data to third parties

Development

Prerequisites

• Node.js 20+
• npm
• Python 3.8+ (for litellm)

Build from Source

# Clone the repository
git clone https://github.com/cruzleedan/vscode-llm-gateway.git
cd vscode-llm-gateway

# Install dependencies
npm install

# Compile TypeScript
npm run compile

# Watch mode (for development)
npm run watch

Testing Locally

# Package the extension
npm run package
# or
npx vsce package

# Install in VS Code
code --install-extension llm-gateway-connector-0.1.0.vsix

# Reload VS Code
# Cmd/Ctrl+Shift+P → "Developer: Reload Window"

Publishing to Marketplace

First-time setup:

1. Create a Publisher account

   • Go to https://marketplace.visualstudio.com/manage
   • Sign in with Microsoft/Azure account
   • Create a new publisher ID

2. Get a Personal Access Token (PAT)

   • Go to https://dev.azure.com
   • User Settings → Personal Access Tokens
   • Create token with Marketplace (Manage) scope
   • Save the token securely

3. Login with vsce

   npx vsce login <your-publisher-id>
   # Paste your PAT when prompted
   

Publishing updates:

# Bump version in package.json first
# Then build and publish
npm run compile
npx vsce package
npx vsce publish

# Or publish directly without packaging
npx vsce publish minor  # Bumps minor version and publishes
npx vsce publish patch  # Bumps patch version and publishes

Debugging

1. Open the project in VS Code
2. Press F5 to launch Extension Development Host
3. Check the Output panel → "LLM Gateway" channel for logs

License

MIT