RAG Pilot

# RAG Pilot [![VS Code Marketplace](https://img.shields.io/badge/VS%20Code-Marketplace-blue)](https://marketplace.visualstudio.com/) [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT) **Supercharge GitHub Copilot with Retrieval-Augmented Generation (RAG)** Give Copilot context from your workspace and any GitHub repository

RAG Pilot is a Visual Studio Code extension that enhances GitHub Copilot by adding semantic search capabilities. Index your codebase and external GitHub repositories, then ask Copilot questions with full context awareness.

✨ Features

🔍 Semantic Search - Find relevant code using AI embeddings, not just keywords
📚 Multi-Source Context - Combine workspace code + external GitHub repos
🤖 GitHub Copilot Integration - Uses your existing Copilot subscription
📍 Source Citations - See which files/repos were used in answers
💾 Persistent Index - Embeddings stored locally for instant retrieval
⚡ Flexible Indexing - Index entire workspace, specific folders, or individual files
🎯 Right-Click Integration - Index files/folders directly from Explorer

🚀 Quick Start

1. Install the Extension

Search for "RAG Pilot" in the VS Code Extensions marketplace and install it.

2. Index Your Code

Option A: Index Entire Workspace

Open Command Palette (Ctrl+Shift+P / Cmd+Shift+P)
Run: RAG Pilot: Index Workspace for Vector Search

Option B: Index Specific Folder

Right-click any folder in Explorer
Select RAG Pilot: Index Specific Folder

Option C: Index Selected Files

Select one or more files in Explorer
Right-click → RAG Pilot: Index Selected Files

3. Add External Repositories (Optional)

Run: RAG Pilot: Add GitHub Repository
Enter repo: owner/repo or full GitHub URL
Wait for download and indexing

4. Ask Questions

Keyboard Shortcut (Fastest):

Press Ctrl+Shift+R (Windows/Linux) or Cmd+Shift+R (Mac)
Type your question

Chat Interface:

Open GitHub Copilot Chat
Type @rag followed by your question
After first use, it stays active (sticky mode)

📖 Usage Examples

@rag how does authentication work in this codebase?

@rag show me examples of API error handling

@rag (after indexing react repo) explain how React hooks are implemented

🛠️ Commands

Command	Description	Shortcut
RAG Pilot: Index Workspace	Index entire workspace	-
RAG Pilot: Index Specific Folder	Index selected folder	Right-click folder
RAG Pilot: Index Selected Files	Index one or more files	Right-click files
RAG Pilot: Add GitHub Repository	Download & index a repo	-
RAG Pilot: List Indexed Sources	View all indexed sources	-
RAG Pilot: Remove Repository	Remove repo from index	-
RAG Pilot: Clear Vector Index	Clear entire index	-
RAG Pilot: Open Chat	Open chat with @rag	`Ctrl+Shift+R`

⚙️ Configuration

{
  // Automatically include RAG context in all chat requests (experimental)
  "copilot-rag.autoAttach": false,
  
  // File patterns to include when indexing (glob patterns)
  "copilot-rag.includePatterns": [
    "**/*.ts", "**/*.js", "**/*.py", "**/*.java", 
    "**/*.cpp", "**/*.go", "**/*.rs", "**/*.md"
  ],
  
  // File patterns to exclude (glob patterns)
  "copilot-rag.excludePatterns": [
    "**/node_modules/**", "**/dist/**", "**/build/**"
  ]
}

🔧 How It Works

Indexing - Code is split into chunks and converted to vector embeddings using all-MiniLM-L6-v2
Storage - Embeddings stored locally in a vector database (Vectra)
Retrieval - Your question is converted to a vector and compared against indexed content
Augmentation - Top 5 most relevant code snippets are retrieved
Generation - Copilot receives your question + context to generate informed answers

📊 Technical Details

Embedding Model: Xenova/all-MiniLM-L6-v2 (runs locally via ONNX)
Vector Store: Vectra (local file-based index)
Chunk Size: 100 lines with 20-line overlap
Retrieval: Top-5 semantic similarity search
Storage Location: ~/.config/Code/User/globalStorage/sudoecho.rag-pilot/

📋 Requirements

Visual Studio Code 1.90.0 or higher
GitHub Copilot extension with active subscription
~200MB disk space (for embedding model and indexes)

🐛 Troubleshooting

"No vector index found"

Run RAG Pilot: Index Workspace or index specific files/folders first

"No Copilot model available"

Ensure GitHub Copilot extension is installed and you're signed in
Check that your Copilot subscription is active

Slow initial indexing

First-time: Downloads embedding model (~90MB) - happens once
Large repositories take time - progress shown in notification

Extension not activating

Reload VS Code window (Ctrl+R / Cmd+R)
Check Output → Extension Host for errors

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Fork the repository
Create your feature branch (git checkout -b feature/amazing-feature)
Commit your changes (git commit -m 'Add some amazing feature')
Push to the branch (git push origin feature/amazing-feature)
Open a Pull Request

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

Built with Xenova/transformers.js
Vector storage by Vectra
Powered by GitHub Copilot and VS Code Language Model API

📧 Support

🐛 Report a bug
💡 Request a feature
⭐ Star this repo if you find it useful!

Made with ❤️ for developers who want smarter AI assistance

sudoecho