WebLLM

Providers

Connect to 18+ AI providers. Configure them in the extension or playground settings.

What are Providers?

Providers are AI service adapters that connect WebLLM to different AI backends. You configure which providers to use, add your API keys, and set priorities.

When a website makes an AI request, WebLLM automatically selects the best provider based on your configuration and the task requirements.

Local AI

Run models directly on your device with zero external data sharing.

Ollama

Local model server

LM Studio

Local models with GUI

Browser (WebGPU)

In-browser inference (optional)

Aggregate Providers

Access hundreds of models through unified gateway APIs.

OpenRouter

100+ models via single API

Portkey

Unified API with load balancing

Model Providers

Direct access to AI model providers with your API keys.

Anthropic

Claude Sonnet, Opus, Haiku

OpenAI

GPT-4o, GPT-4, GPT-3.5 Turbo

Google Generative AI

Gemini models

Google Vertex AI

Enterprise Gemini on GCP

Anthropic Vertex AI

Claude on Google Cloud

Azure OpenAI

GPT models on Azure

Mistral AI

Mistral, Mixtral models

Cohere

Command models

DeepSeek

Fast, budget-friendly models

Groq

Ultra-low latency inference

Fireworks AI

Fast model inference

Together.ai

Open source models

OS and Browser Providers

Native inference engines from operating systems and browsers (expected soon).

Chrome Built-in AI

Coming Soon

Coming soon

Windows Copilot Runtime

Coming Soon

Coming soon

macOS Intelligence

Coming Soon

Coming soon

Sponsored Gateways

Gateways provided by developers as a convenience to make WebLLM available to all users, including mobile and those who haven't set up the extension yet.

Developer Gateways

Your hosted WebLLM endpoints

Learn more →

How to Configure Providers

1. Add API Keys

Go to the Providers page in the extension or playground. Click on a provider and enter your API key from their website.

2. Set Priorities

Drag providers to reorder them. Higher priority providers are preferred when routing requests.

3. Enable/Disable

Toggle providers on/off. Disabled providers won't be used even if they have API keys configured.

How WebLLM Selects Providers

WebLLM automatically selects the best provider for each request based on task requirements, model capabilities, and your priorities. The intelligent routing system scores providers by 16 criteria including speed, quality, cost, and capabilities.

Learn more about model routing →

Next Steps

Configure Your Providers

Set up API keys and priorities in the playground

Go to Providers →
Learn About Routing

Understand how WebLLM selects models

View Routing Docs →