Run GPT model on the browser with WebGPU. An implementation of GPT inference in less than ~1500 lines of vanilla Javascript.

An amazing UI for OpenAI's ChatGPT (Website + Windows + MacOS + Linux)

User-friendly AI Interface (Supports Ollama, OpenAI API, ...)

WebGPU LLM inference tuned by hand

Universal LLM Deployment Engine with ML Compilation

Companion application for Video DownloadHelper browser add-on

WebGPU Samples