PairIt

PairIt – Your Code. Your Browser. Your AI.

PairIt is a Chrome extension and Python SDK that connects local Python applications to browser-based AI sessions and local AI models running inside Chrome. Why PairIt? PairIt lets developers use AI from Python without building separate integrations for every provider. You can connect to an active ChatGPT, Claude, or Gemini browser session, or run supported chat, vision, and embedding models locally in Chrome using WebGPU or WASM. What PairIt provides • ChatGPT, Claude, and Gemini browser-session support • Local chat and coding models • Local vision and image-processing models • Local text and multilingual embedding models • WebGPU acceleration with WASM fallback • Model downloads from Hugging Face • SHA-256 verification for downloaded model files • Local model storage using Chrome Cache Storage and IndexedDB • Real-time response streaming for browser sessions • A local WebSocket bridge at ws://127.0.0.1:8765 • Extension popup and Chrome side-panel support • Explicit ON/OFF control over the local connection • No PairIt cloud service or PairIt telemetry How it works The PairIt Python SDK starts a local bridge on your computer. The Chrome extension connects to that bridge and routes requests to either an active browser AI session or a local model running inside Chrome. Browser-session requests are handled by the selected provider's website using the browser session you are already signed into. Local model requests are processed inside Chrome after the required model files have been downloaded and verified. How to use PairIt 1. Install the Python SDK with pip. 2. Install and pin the PairIt Chrome extension. 3. Open ChatGPT, Claude, or Gemini in Chrome and sign in, or choose a local model from the extension. 4. Turn PairIt ON from the extension popup. 5. Create a Python client with the model or provider you want to use. 6. Send chat, vision, or embedding requests through the PairIt SDK. PairIt is designed for developers who want a simple Python interface for browser AI sessions and private, local in-browser inference without managing multiple provider integrations.