How is this AI private?

The AI model is downloaded and run entirely in your browser's memory using WebGPU. None of your chat messages are ever sent to an external server.

Why does it take so long to load the first time?

The tool needs to download the AI model weights (which can be over 1GB depending on the model) to your device. Once downloaded, it is cached in your browser for faster loading next time.

Private AI Chat — ASAPUtils

What is Private AI Chat?

Private AI Chat is an entirely on-device, fully private AI assistant powered by @mlc-ai/web-llm. It runs Large Language Models (LLMs) directly in your browser using WebGPU.

Features

100% Private: Your data never leaves your device. No API keys, no cloud processing, no tracking.
Offline Capable: Once the model is downloaded and cached, you can chat with the AI without an internet connection.
Powered by WebGPU: Leverages your device’s graphics card to run models efficiently in the browser.

How to Use

Initialize the Model: Click the “Load Engine” button to initialize the AI.
Wait for Download: The first time you load a model, it will download the necessary files and cache them in your browser. This can take a few minutes depending on your internet connection.
Start Chatting: Once loaded, simply type your message and click “Send”.

Note: This tool requires a modern browser with WebGPU support (like Google Chrome, Microsoft Edge, or Firefox).