Model Zoo | ADI Protocol

Welcome to Model Zoo

Run powerful AI models entirely in your browser

One-Time Download

When you select a model, it will be downloaded directly to your browser. This may take a few minutes depending on your connection, but it only happens once — models are cached for future visits.

Runs on Your GPU

Models run locally using WebGPU, which leverages your device's graphics card for fast inference. No server required — everything happens right on your machine.

100% Private

Your conversations never leave your device. There's no server processing, no data collection, and no internet required after the initial model download.

Instant After Loading

Once a model is loaded into memory, responses generate at blazing speeds directly on your hardware. Enjoy real-time AI conversations offline.

Requirements

Modern browser with WebGPU support (Chrome 113+, Edge 113+)
GPU with at least 4GB VRAM recommended
Sufficient storage space for model files (0.7GB - 2.2GB)

Navigation

AI Chat Chat History 0 How It Works

Select Model

Llama 3.2 1B 0.7 GB

Llama 3.2 3B 1.8 GB

Phi 3.5 Mini 2.2 GB

Gemma 2 2B 1.4 GB

Qwen 2.5 1.5B 1.0 GB

SmolLM2 1.7B 1.1 GB

Quick Actions

Clear Chat

Export Chat

Model: None

Tokens: 0

Speed: -

Select a Model to Begin

Choose an AI model from the sidebar or the cards below. Models will be downloaded to your browser on first use.

Llama 3.2 1B

0.7 GB

Fast

Llama 3.2 3B

1.8 GB

Balanced

Phi 3.5 Mini

2.2 GB

Balanced

Gemma 2 2B

1.4 GB

Fast

Initializing model...