WebLLM

Documentation

Everything you need to build AI-powered applications with WebLLM

Quick Start
Get up and running in 5 minutes

Install WebLLM and make your first AI request.

Get Started →
API Reference
Complete API documentation

Explore all available methods and options.

View API →
Guides
Step-by-step tutorials

Learn how to integrate WebLLM in your projects.

Browse Guides →
Model Routing
Intelligent model selection

Learn how WebLLM automatically selects the best model for each request.

Learn More →
Gateways
Mobile & fallback support

Deploy WebLLM to mobile devices and provide fallback support.

Learn More →
Vercel AI SDK
Integration with Vercel AI

Use WebLLM with the Vercel AI SDK.

Learn More →

What is WebLLM?

WebLLM is an open protocol for AI inference designed as a web standard. It provides a standardized interface for accessing AI models directly from the browser, similar to how localStorage or geolocation work.

Key features:

  • Browser-native AI access via Chrome extension
  • Support for 14+ providers (OpenAI, Anthropic, Google, local models, etc.)
  • Zero infrastructure costs - users bring their own AI
  • Mobile support via Developer Gateways
  • Open source (MIT licensed)

Installation

npm install webllm

See the installation guide for more details.