VSCode AI Helper

VSCode AI Helper is a Visual Studio Code extension project that brings an AI-powered chat assistant right into your editor. It connects to a HuggingFace model via API, letting you chat with your own LLM, Copilot-style.

Note: This project is for demonstration and learning purposes. The default HuggingFace model endpoint is currently inactive ("sleeping"). Feel free to fork, edit, and connect your own model!

---

✨ Features

• AI Chat Panel
  Interact with a conversational AI model directly inside VSCode.
• Contextual Assistance
  Get code explanations, suggestions, bug fixes, or generate snippets.
• Customizable Backend
  Designed to connect with any HuggingFace-compatible text-generation API—just update the endpoint!
• Private & Local
  No code or chat data leaves your machine except requests sent to your chosen model endpoint.

---

🧐 What Is This?

This is a VSCode extension project that creates a sidebar panel for chatting with an AI assistant, similar to Copilot Chat, but powered by your own HuggingFace model.

• All AI responses are fetched from a HuggingFace API endpoint defined in the extension settings.
• The extension comes pre-configured for demonstration, but the example endpoint is currently inactive.
• To use it, edit the project and connect your own model endpoint.

---

🛠️ Project Structure

vscode-ai-helper-project/
├── src/            # Extension source code
├── assets/         # Images and demo assets
├── package.json    # Extension manifest
└── README.md       # (You are here!)

---

🧑‍💻 How To Connect Your Own Model

Want to make this work with your own HuggingFace or other LLM endpoint?

1. Host Your Model:
   Deploy a text-generation model (like Llama, Mistral, Falcon, etc.) on HuggingFace Spaces or your own server.

2. Find Your Endpoint:
   Get the API endpoint URL and (if needed) API Key.

3. Edit the Code:

   • Open the extension source (look for API calls in the code, typically in src/).
   • Replace the default endpoint with your own.
   • Optionally, add authentication headers or other needed config.

4. (Optional) Build & Package:

   • If you want to test in VSCode, run npm install && npm run build inside the project directory, then load it as an unpacked extension.

---

⚙️ Settings Reference

Setting Name	Description
ai-helper.huggingfaceEndpoint	URL to your HuggingFace inference API
ai-helper.apiKey	HuggingFace API token (if required)
ai-helper.defaultPromptTemplate	Optional: Customize the prompt template sent to your model

---

📸 Demo

---

❓ FAQ

Q: Why does the extension not respond?
A: The demo HuggingFace endpoint is currently sleeping/inactive. To use this, connect your own model as described above.

Q: Can I use any LLM?
A: Yes—just point the extension at any text-generation endpoint with a compatible API.

Q: Does this share my code with anyone?
A: Only with the endpoint you configure in the settings. You control where your data goes!

---

🏗️ For Developers

• Fork the repo, customize the extension, and make it your own!
• Contributions and PRs welcome.

---

📄 License

MIT License

---

Build your own AI Copilot for VSCode—fully under your control!