colabpilot 0.1.1

Bridge Google Colab GPU to LLM chat UIs (ChatGPT, Claude, Gemini)

ColabPilot

Bridge Google Colab GPU to LLM chat UIs (ChatGPT, Claude, Gemini).

Setup

In any Google Colab cell:

!pip install colabpilot
import colabpilot
colabpilot.connect()

Then install the ColabPilot Chrome Extension and open any LLM chat. Code blocks will get a "Run in Colab" button that executes on your Colab GPU and sends results back.

How it works

1. colabpilot.connect() registers Python callbacks with Colab's built-in JS-to-Python RPC system
2. The Chrome extension detects the bridge via a DOM marker
3. When you click "Run in Colab" on a code block, the extension calls the Python callback
4. Code executes in a shared namespace (variables persist across runs)
5. stdout, stderr, and errors are captured and sent back to the chat

Requirements

• Google Colab (free tier works)
• ColabPilot Chrome Extension
• Any supported LLM chat: ChatGPT, Claude, or Gemini