🚀 InfiniAI Provider for Copilot

Welcome to InfiniAI Provider for Copilot! This is a model provider extension designed specifically for VS Code Copilot. With this extension, you can seamlessly integrate the powerful model gateway capabilities of InfiniAI into VS Code Copilot, giving you the freedom to use top-tier AI models.

💡 Usage

Just a few simple steps to start your InfiniAI journey:

1. 📥 Install Extension: Click here to install the extension.
2. 💬 Open Copilot: Open the GitHub Copilot Chat interface in VS Code.
3. ⚙️ Manage Models: Click the model picker below the chat input box and select "Manage Models...".
4. ✅ Select InfiniAI: Click "Add Models" and then select the "InfiniAI" provider.
5. 🧭 Select Plan (first time only): If you haven't configured infiniai.plan yet, you'll be prompted to pick Standard Plan or Coding Plan. The choice will be saved to your VS Code user settings.
6. 🔑 Configure Key: Enter your InfiniAI API Key for the selected plan (the key is stored securely in VS Code Secret Storage).
7. 🎯 Pick Models: Select the specific models you wish to use in the model picker.

ℹ️ Extension Information

• Name: InfiniAI Provider for Copilot
• Version: See package.json

✅ Prerequisites

Before you begin, please ensure you meet the following requirements:

• 💻 VS Code Version: >= 1.104.0
• 🧩 Copilot Extension: github.copilot-chat extension installed
• 🔑 API Key: A valid InfiniAI API Key (get it from infiniai.ai)
• 🟢 Node.js: (Required only for development and building)

🛠️ Installation & Build (Development Guide)

If you are a developer and want to build or modify this project yourself:

1. Install Dependencies

npm install

2. Compile TypeScript

npm run compile

3. Package VSIX (Optional)

npm run build

🐛 Run in Extension Development Host

1. Open this repository in VS Code.
2. Press F5 to launch the Extension Development Host.
3. In the development host, open Copilot Chat; you should be able to see and use the InfiniAI Provider.

📝 Activation & Logging

• Activation Events: The extension activates when events declared in package.json are triggered (e.g., onStartupFinished or when running a command).
• View Logs:
  1. Open the Output Panel (View → Output or Ctrl+Shift+U).
  2. Select the InfiniAI channel from the dropdown menu in the top right corner.

⚙️ Configuration (Common)

You can adjust the following parameters in VS Code Settings:

• infiniai.plan: Select the InfiniAI billing plan — "standard" (pay-per-token) or "coding" (Coding Plan subscription). If unset, the extension treats it as "standard" for API routing, and will prompt you to pick a plan the next time it needs to ask for an API key interactively.
• infiniai.baseUrl: Base URL for OpenAI-compatible API, Standard Plan (Default: https://cloud.infini-ai.com/maas/v1).
• infiniai.anthropic.baseUrl: Base URL for Anthropic-compatible API, Standard Plan (Default: https://cloud.infini-ai.com/maas).
• infiniai.coding.baseUrl: Base URL for OpenAI-compatible API, Coding Plan (Default: https://cloud.infini-ai.com/maas/coding/v1).
• infiniai.coding.anthropic.baseUrl: Base URL for Anthropic-compatible API, Coding Plan (Default: https://cloud.infini-ai.com/maas/coding).
• infiniai.imageInputModels: Force-enable image input for specific model IDs (supports * wildcard, e.g. kimi-*). Useful when a multimodal model doesn’t include a name marker like -vision.
• infiniai.disableImageInputModels: Force-disable image input for specific model IDs (supports * wildcard).
• infiniai.retry: Request retry policy (enabled, max attempts, interval in ms).
• infiniai.delay: Fixed delay between requests (in milliseconds).

🧠 Thinking / Reasoning Parameters

Status (v0.2.1+): Thinking/reasoning parameters are currently not sent to the API. The extension recognizes these model configuration options internally (enable_thinking, thinking_budget, reasoning_effort, thinking), but the code to include them in API requests is commented out. This means:

• Models that support thinking/reasoning will use their default behavior
• You cannot currently enable or configure thinking through this extension
• Thinking content received from models is tracked internally but not displayed in Copilot Chat

If you need thinking capabilities enabled, this would require modifying the extension code in src/openai/openaiApi.ts to uncomment the relevant parameter handling.

⌨️ Commands

• infiniai.setApikey: Run this command via the Command Palette (Ctrl+Shift+P) to set or update your InfiniAI API Key. You will be prompted to choose which plan (Standard or Coding) to configure the key for.

🧪 Tests

npm test

🔍 Debugging Tips

If the extension does not activate or shows no logs:

• 🧐 Ensure you are viewing the Extension Development Host window.
• 📄 Check the InfiniAI channel in the Output Panel.
• 🐞 Open Developer Tools (Help → Toggle Developer Tools) to check for console errors.
• 🔄 Try Reload Window (Developer: Reload Window).
• 📁 Confirm that the out/extension.js file exists (ensure you have run npm run compile).

🤝 Contributing & Feedback

We welcome your participation!

• 🐛 Submit Issues: GitHub Issues
• 🔀 Contribute Code: Feel free to Fork this repository and submit a Pull Request.

📄 License

MIT License