This is an agent-based GitHub Copilot Extension that provides a way to interact with GitHub Models, directly in Copilot Chat.

You can ask it things like "what models are available?" or "which model should I use for my use-case?". You can also ask it to execute a basic prompt using a particular model.

Description	Image
User asking @models for a recommended model that is low-cost and supports function calling
Dialogue with the Extension, asking for an OpenAI model
User executing a basic prompt using the model that the extension recommended

1. Install dependencies:

npm install

2. Run the server

• To run in development mode:

npm run dev

• To build and run in production mode:

npm run build && npm start

3. Follow this guide to make your server accessible to the internet

In short, we would expose a public URL for our local server using the following command (follow the guide for detailed setup instructions):

ngrok http https://localhost:3000

4. Follow this guide to create a GitHub app
5. Follow this guide to configure the app to use the public URL from (3.)
6. Use your newly installed app! On any copilot enabled page, type @your-app-name <prompt> (from the app created in 4.) to interact with your local installation of this extension!

• Using Copilot Extensions
• About building Copilot Extensions
• Set up process
• Communicating with the Copilot platform
• Communicating with GitHub