Free AI for Copilot Chat

Use Free AI models from the Copilot Chat model picker, with zero proxy setup and full BYOK control.

This extension plugs an OpenAI-compatible API into VS Code Copilot Chat. It is preconfigured for free-ai, but you can also point it at another compatible baseUrl and define your own model list.

Features

• Use Free AI models directly from the Copilot Chat model picker
• Configure baseUrl and apiKey
• Store API keys in VS Code SecretStorage
• Expose any upstream model IDs you want through free-ai-copilot.models
• Supports streaming, tool calling, and image input flags
• Preserves Copilot agent mode and tool integrations

Getting Started

1. Install the extension
2. Run Free AI: Set API Key
3. Open settings and review:
   • free-ai-copilot.baseUrl
   • free-ai-copilot.models
4. Open Copilot Chat and pick one of your configured Free AI models

Example free-ai-copilot.models entry:

[
  {
    "id": "gpt-5.4",
    "name": "Free AI GPT-5.4",
    "family": "free-ai",
    "version": "gpt-5.4",
    "detail": "GPT-5.4 default model",
    "maxInputTokens": 1048576,
    "maxOutputTokens": 131072,
    "capabilities": {
      "toolCalling": true,
      "imageInput": true,
      "thinking": true
    }
  }
]

Default Endpoint

• https://api.freetheai.xyz/v1

Settings

Notes

• The default bundled model is gpt-5.4. If your upstream endpoint uses a different model ID, change it to the exact value returned by /models.
• Use Free AI: Open Project Page if you need the upstream project and key instructions.

License

MIT