> ## Documentation Index
> Fetch the complete documentation index at: https://mcp-b-sync-npm-packages-docs-bf03420.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.
# AI Browsers
> Native browser AI assistants that support WebMCP tools out of the box.
AI browsers are web browsers with built-in AI assistants that can interact with websites using WebMCP. When you register tools via `navigator.modelContext`, these browser agents can discover and use them automatically.
## Supported Browsers
Google Chrome's built-in AI assistant (coming soon)
AI-native browser with search integration
Browser with AI features and WebMCP support
Microsoft Edge with Copilot integration
Browser AI support is actively being developed. Check individual browser documentation for current WebMCP compatibility.
## How It Works
AI browsers have native access to the `navigator.modelContext` API:
```mermaid theme={null}
graph LR
A[Browser AI] -->|navigator.modelContext| B[Your Website]
B -->|Registered Tools| A
A -->|Tool Calls| B
B -->|Results| A
```
When a user asks the browser AI to perform an action:
1. The AI discovers tools registered on the current page
2. It selects appropriate tools based on the user's request
3. Tools execute in the browser with the user's session
4. Results are returned to the AI for response generation
## Preparing Your Website
To ensure your website works with AI browsers:
Use `navigator.modelContext.registerTool()` to expose functionality:
```typescript theme={null}
navigator.modelContext.registerTool({
name: 'search_products',
description: 'Search the product catalog',
inputSchema: {
type: 'object',
properties: {
query: { type: 'string' }
}
},
async execute({ query }) {
const results = await searchProducts(query);
return { content: [{ type: 'text', text: JSON.stringify(results) }] };
}
});
```
Browser AIs rely on tool descriptions to understand when to use them. Be specific:
```typescript theme={null}
// Good
description: 'Add a product to the shopping cart by product ID and quantity'
// Too vague
description: 'Add to cart'
```
Browser AIs run with user permissions. Ensure your tools validate inputs:
```typescript theme={null}
async execute({ productId, quantity }) {
if (quantity < 1 || quantity > 100) {
throw new Error('Quantity must be between 1 and 100');
}
// ... rest of implementation
}
```
## Testing Without AI Browsers
While waiting for AI browser support, test your tools with:
Add a drop-in AI assistant to your site
Test tools via the browser extension
Automated testing with MCP clients
## Security Considerations
AI browser tools execute with the same permissions as your website. They can access localStorage, cookies, and make authenticated requests.
* Tools inherit the user's browser session
* Same-origin policy applies
* Browser AI will request user permission for sensitive operations
Review comprehensive security guidelines