🤖
Browser AI Chat
Chat with an AI that runs entirely in your browser using WebLLM. No data leaves your device — fully private, powered by WebGPU.
AIaichatllmwebllm🔒 Browser-only
📋 How to use Browser AI Chat
- 1Click 'Load Model' and choose Llama 3.2 (fast) or Phi-3.5 (smarter)
- 2Wait for the one-time model download (cached for future sessions)
- 3Type your question and press Enter to chat
- 4All processing happens in your browser — nothing is sent to any server
Try it now
Choose a model
Runs entirely in your browser · Downloaded once · Cached for future sessions
Requires a WebGPU-capable GPU. First load downloads the model to your browser cache.
Examples
Quick question
Input: "How do I merge two PDFs?"
Output: "AI explains how to use the PDF Merger tool"
General help
Input: "What tools are available for images?"
Output: "AI lists image tools with descriptions"
Frequently Asked Questions
Does this use the internet?▾
Only for the initial model download. After that, everything runs in your browser with no internet required.
What are the system requirements?▾
A WebGPU-capable browser (Chrome 113+, Edge 113+) and a modern GPU. 2–8 GB of GPU memory depending on the model.
Is my data private?▾
Yes. Your messages never leave your device. The AI runs entirely on your hardware.
Which model should I choose?▾
Llama 3.2 1B is fast and lightweight (0.7 GB). Phi-3.5 Mini gives better reasoning but takes longer to download (2.4 GB).