# Scrapfly Documentation
## Table of Contents
### Dashboard
- [Intro](https://scrapfly.io/docs)
- [Project](https://scrapfly.io/docs/project)
- [Account](https://scrapfly.io/docs/account)
- [Workspace & Team](https://scrapfly.io/docs/workspace-and-team)
- [Billing](https://scrapfly.io/docs/billing)
### Products
#### MCP Server
- [Getting Started](https://scrapfly.io/docs/mcp/getting-started)
- [Tools & API Spec](https://scrapfly.io/docs/mcp/tools)
- [Authentication](https://scrapfly.io/docs/mcp/authentication)
- [Examples & Use Cases](https://scrapfly.io/docs/mcp/examples)
- [FAQ](https://scrapfly.io/docs/mcp/faq)
##### Integrations
- [Overview](https://scrapfly.io/docs/mcp/integrations)
- [Claude Desktop](https://scrapfly.io/docs/mcp/integrations/claude-desktop)
- [Claude Code](https://scrapfly.io/docs/mcp/integrations/claude-code)
- [ChatGPT](https://scrapfly.io/docs/mcp/integrations/chatgpt)
- [Cursor](https://scrapfly.io/docs/mcp/integrations/cursor)
- [Cline](https://scrapfly.io/docs/mcp/integrations/cline)
- [Windsurf](https://scrapfly.io/docs/mcp/integrations/windsurf)
- [Zed](https://scrapfly.io/docs/mcp/integrations/zed)
- [Roo Code](https://scrapfly.io/docs/mcp/integrations/roo-code)
- [VS Code](https://scrapfly.io/docs/mcp/integrations/vscode)
- [LangChain](https://scrapfly.io/docs/mcp/integrations/langchain)
- [LlamaIndex](https://scrapfly.io/docs/mcp/integrations/llamaindex)
- [CrewAI](https://scrapfly.io/docs/mcp/integrations/crewai)
- [OpenAI](https://scrapfly.io/docs/mcp/integrations/openai)
- [n8n](https://scrapfly.io/docs/mcp/integrations/n8n)
- [Make](https://scrapfly.io/docs/mcp/integrations/make)
- [Zapier](https://scrapfly.io/docs/mcp/integrations/zapier)
- [Vapi AI](https://scrapfly.io/docs/mcp/integrations/vapi)
- [Agent Builder](https://scrapfly.io/docs/mcp/integrations/agent-builder)
- [Custom Client](https://scrapfly.io/docs/mcp/integrations/custom-client)
#### Web Scraping API
- [Getting Started](https://scrapfly.io/docs/scrape-api/getting-started)
- [API Specification]()
- [Monitoring](https://scrapfly.io/docs/monitoring)
- [Customize Request](https://scrapfly.io/docs/scrape-api/custom)
- [Debug](https://scrapfly.io/docs/scrape-api/debug)
- [Anti Scraping Protection](https://scrapfly.io/docs/scrape-api/anti-scraping-protection)
- [Proxy](https://scrapfly.io/docs/scrape-api/proxy)
- [Proxy Mode](https://scrapfly.io/docs/scrape-api/proxy-mode)
- [Proxy Mode - Screaming Frog](https://scrapfly.io/docs/scrape-api/proxy-mode/screaming-frog)
- [Proxy Mode - Apify](https://scrapfly.io/docs/scrape-api/proxy-mode/apify)
- [(Auto) Data Extraction](https://scrapfly.io/docs/scrape-api/extraction)
- [Javascript Rendering](https://scrapfly.io/docs/scrape-api/javascript-rendering)
- [Javascript Scenario](https://scrapfly.io/docs/scrape-api/javascript-scenario)
- [SSL](https://scrapfly.io/docs/scrape-api/ssl)
- [DNS](https://scrapfly.io/docs/scrape-api/dns)
- [Cache](https://scrapfly.io/docs/scrape-api/cache)
- [Batch (Multi-URL Scraping)](https://scrapfly.io/docs/scrape-api/batch)
- [Session](https://scrapfly.io/docs/scrape-api/session)
- [Webhook](https://scrapfly.io/docs/scrape-api/webhook)
- [Schedule](https://scrapfly.io/docs/scrape-api/schedule)
- [Screenshot](https://scrapfly.io/docs/scrape-api/screenshot)
- [Errors](https://scrapfly.io/docs/scrape-api/errors)
- [Timeout](https://scrapfly.io/docs/scrape-api/understand-timeout)
- [Throttling](https://scrapfly.io/docs/throttling)
- [Troubleshoot](https://scrapfly.io/docs/scrape-api/troubleshoot)
- [Billing](https://scrapfly.io/docs/scrape-api/billing)
- [FAQ](https://scrapfly.io/docs/scrape-api/faq)
#### Crawler API
- [Getting Started](https://scrapfly.io/docs/crawler-api/getting-started)
- [API Specification]()
- [Retrieving Results](https://scrapfly.io/docs/crawler-api/results)
- [WARC Format](https://scrapfly.io/docs/crawler-api/warc-format)
- [Data Extraction](https://scrapfly.io/docs/crawler-api/extraction-rules)
- [Webhook](https://scrapfly.io/docs/crawler-api/webhook)
- [Schedule](https://scrapfly.io/docs/crawler-api/schedule)
- [Billing](https://scrapfly.io/docs/crawler-api/billing)
- [Errors](https://scrapfly.io/docs/crawler-api/errors)
- [Troubleshoot](https://scrapfly.io/docs/crawler-api/troubleshoot)
- [FAQ](https://scrapfly.io/docs/crawler-api/faq)
#### Screenshot API
- [Getting Started](https://scrapfly.io/docs/screenshot-api/getting-started)
- [API Specification]()
- [Accessibility Testing](https://scrapfly.io/docs/screenshot-api/accessibility)
- [Webhook](https://scrapfly.io/docs/screenshot-api/webhook)
- [Schedule](https://scrapfly.io/docs/screenshot-api/schedule)
- [Billing](https://scrapfly.io/docs/screenshot-api/billing)
- [Errors](https://scrapfly.io/docs/screenshot-api/errors)
#### Extraction API
- [Getting Started](https://scrapfly.io/docs/extraction-api/getting-started)
- [API Specification]()
- [Rules Template](https://scrapfly.io/docs/extraction-api/rules-and-template)
- [LLM Extraction](https://scrapfly.io/docs/extraction-api/llm-prompt)
- [AI Auto Extraction](https://scrapfly.io/docs/extraction-api/automatic-ai)
- [Webhook](https://scrapfly.io/docs/extraction-api/webhook)
- [Billing](https://scrapfly.io/docs/extraction-api/billing)
- [Errors](https://scrapfly.io/docs/extraction-api/errors)
- [FAQ](https://scrapfly.io/docs/extraction-api/faq)
#### Proxy Saver
- [Getting Started](https://scrapfly.io/docs/proxy-saver/getting-started)
- [Fingerprints](https://scrapfly.io/docs/proxy-saver/fingerprints)
- [Optimizations](https://scrapfly.io/docs/proxy-saver/optimizations)
- [SSL Certificates](https://scrapfly.io/docs/proxy-saver/certificates)
- [Protocols](https://scrapfly.io/docs/proxy-saver/protocols)
- [Pacfile](https://scrapfly.io/docs/proxy-saver/pacfile)
- [Secure Credentials](https://scrapfly.io/docs/proxy-saver/security)
- [Billing](https://scrapfly.io/docs/proxy-saver/billing)
#### Cloud Browser API
- [Getting Started](https://scrapfly.io/docs/cloud-browser-api/getting-started)
- [Proxy & Geo-Targeting](https://scrapfly.io/docs/cloud-browser-api/proxy)
- [Unblock API](https://scrapfly.io/docs/cloud-browser-api/unblock)
- [Captcha Solver](https://scrapfly.io/docs/cloud-browser-api/captcha-solver)
- [File Downloads](https://scrapfly.io/docs/cloud-browser-api/file-downloads)
- [Session Resume](https://scrapfly.io/docs/cloud-browser-api/session-resume)
- [Human-in-the-Loop](https://scrapfly.io/docs/cloud-browser-api/human-in-the-loop)
- [Debug Mode](https://scrapfly.io/docs/cloud-browser-api/debug-mode)
- [Bring Your Own Proxy](https://scrapfly.io/docs/cloud-browser-api/bring-your-own-proxy)
- [Browser Extensions](https://scrapfly.io/docs/cloud-browser-api/extensions)
- [Native Browser MCP](https://scrapfly.io/docs/cloud-browser-api/mcp)
- [DevTools Protocol](https://scrapfly.io/docs/cloud-browser-api/cdp-reference)
##### Integrations
- [Puppeteer](https://scrapfly.io/docs/cloud-browser-api/puppeteer)
- [Playwright](https://scrapfly.io/docs/cloud-browser-api/playwright)
- [Selenium](https://scrapfly.io/docs/cloud-browser-api/selenium)
- [Vercel Agent Browser](https://scrapfly.io/docs/cloud-browser-api/agent-browser)
- [Browser Use](https://scrapfly.io/docs/cloud-browser-api/browser-use)
- [Stagehand](https://scrapfly.io/docs/cloud-browser-api/stagehand)
- [Billing](https://scrapfly.io/docs/cloud-browser-api/billing)
- [Errors](https://scrapfly.io/docs/cloud-browser-api/errors)
### Tools
- [Antibot Detector](https://scrapfly.io/docs/tools/antibot-detector)
### SDK
- [Golang](https://scrapfly.io/docs/sdk/golang)
- [Python](https://scrapfly.io/docs/sdk/python)
- [Rust](https://scrapfly.io/docs/sdk/rust)
- [TypeScript](https://scrapfly.io/docs/sdk/typescript)
- [Scrapy](https://scrapfly.io/docs/sdk/scrapy)
### Integrations
- [Getting Started](https://scrapfly.io/docs/integration/getting-started)
- [LangChain](https://scrapfly.io/docs/integration/langchain)
- [LlamaIndex](https://scrapfly.io/docs/integration/llamaindex)
- [CrewAI](https://scrapfly.io/docs/integration/crewai)
- [Zapier](https://scrapfly.io/docs/integration/zapier)
- [Make](https://scrapfly.io/docs/integration/make)
- [n8n](https://scrapfly.io/docs/integration/n8n)
### Academy
- [Overview](https://scrapfly.io/academy)
- [Web Scraping Overview](https://scrapfly.io/academy/scraping-overview)
- [Tools](https://scrapfly.io/academy/tools-overview)
- [Reverse Engineering](https://scrapfly.io/academy/reverse-engineering)
- [Static Scraping](https://scrapfly.io/academy/static-scraping)
- [HTML Parsing](https://scrapfly.io/academy/html-parsing)
- [Dynamic Scraping](https://scrapfly.io/academy/dynamic-scraping)
- [Hidden API Scraping](https://scrapfly.io/academy/hidden-api-scraping)
- [Headless Browsers](https://scrapfly.io/academy/headless-browsers)
- [Hidden Web Data](https://scrapfly.io/academy/hidden-web-data)
- [JSON Parsing](https://scrapfly.io/academy/json-parsing)
- [Data Processing](https://scrapfly.io/academy/data-processing)
- [Scaling](https://scrapfly.io/academy/scaling)
- [Walkthrough Summary](https://scrapfly.io/academy/walkthrough-summary)
- [Scraper Blocking](https://scrapfly.io/academy/scraper-blocking)
- [Proxies](https://scrapfly.io/academy/proxies)
---
# Native Browser MCP
Scrapium natively integrates MCP: the browser itself is the MCP server. Websites register tools via `navigator.modelContext`, and the browser exposes them through the WebMCP CDP domain. Enable it with `enable_mcp=true` to get a streamable-HTTP MCP endpoint that any AI agent can connect to for structured tool discovery and invocation. No external server required.
1. [Cloud Browser](https://scrapfly.io/docs/cloud-browser-api/getting-started)
2. Native Browser MCP
Scrapium natively integrates the [Model Context Protocol](https://modelcontextprotocol.io/) (MCP). **The browser itself is the MCP server.** Websites register tools via the `navigator.modelContext` API, and the browser exposes them through the [WebMCP CDP domain](https://scrapfly.io/docs/cloud-browser-api/cdp-reference/WebMCP). When enabled on a Cloud Browser session, Scrapfly exposes a streamable-HTTP MCP endpoint that AI agents can connect to directly, no external server required.
## What is WebMCP?
[WebMCP](https://developer.chrome.com/blog/webmcp-epp) is a proposed web standard developed jointly by Google and Microsoft under the [W3C Web Machine Learning Community Group](https://www.w3.org/groups/cg/webmachinelearning/). It lets websites expose structured, callable **tools** directly to AI agents through the browser's `navigator.modelContext` API.
Instead of clicking buttons and parsing screenshots, AI agents get a structured contract:
- **Tool discovery** - agents learn what actions a page supports (add to cart, search, fill form, etc.)
- **Typed inputs/outputs** - each tool declares a JSON schema for its parameters and return values
- **Reliable execution** - tool calls go through the browser's native API, not fragile DOM selectors
**WebMCP vs MCP:** WebMCP operates **client-side within the browser** (via `navigator.modelContext`), while the broader [Model Context Protocol](https://modelcontextprotocol.io/) is a back-end protocol connecting AI platforms to service providers through hosted servers. They share the same wire format but serve different layers of the stack.
## How It Works with Cloud Browser
When you enable MCP on a Cloud Browser session, the Scrapfly MCP Server connects to the browser via CDP (Chrome DevTools Protocol) and registers two types of tools:
1. **Browser interaction tools**, human-like actions (clickOn, fill, typeText, scroll, pressKey, hover, selectOption, etc.) that call [Antibot CDP commands](https://scrapfly.io/docs/cloud-browser-api/cdp-reference/Antibot) directly. These are always available and use realistic mouse/keyboard timing.
2. **Page-registered WebMCP tools**, if the web page supports [WebMCP](https://developer.chrome.com/blog/webmcp-epp), it can register custom tools (e.g. `addToCart`, `searchProducts`) via `navigator.modelContext.registerTool()`. These are discovered automatically and invoked via the `WebMCP.invokeTool` CDP command.
3. **Native CDP tools**, screenshot, evaluate JavaScript, get accessibility snapshot, page URL, and performance metrics, all via CDP.
- [ Via Scrapfly MCP Server ](#scrapfly-mcp-panel)
- [ Direct Browser (chrome-devtools-mcp) ](#direct-mcp-panel)
The Scrapfly MCP Server manages the browser session and proxies all tool calls through a single MCP connection. Browser actions use human-like [Antibot CDP commands](https://scrapfly.io/docs/cloud-browser-api/cdp-reference/Antibot) with realistic mouse/keyboard timing. Any MCP-compatible AI agent works out of the box.
Scrapium exposes a **streamable-HTTP MCP endpoint** at `/mcp` that wraps the CDP WebSocket into the MCP protocol, the same approach as [chrome-devtools-mcp](https://github.com/GoogleChromeLabs/chrome-devtools-mcp) (Node.js) but built in Go and running alongside the browser. Any MCP client can connect directly to this endpoint. Browser interactions use [Antibot CDP commands](https://scrapfly.io/docs/cloud-browser-api/cdp-reference/Antibot) for human-like timing.
## WebMCP CDP Domain
The browser exposes registered MCP tools through the [**WebMCP**](https://scrapfly.io/docs/cloud-browser-api/cdp-reference/WebMCP) CDP domain. This is the low-level interface that powers the streamable-HTTP MCP endpoint.
| Symbol | Type | Description |
|---|---|---|
| `WebMCP.enable` | Command | Start receiving tool events. Fires an initial `toolsAdded` event with all available Antibot tools and their JSON schemas. |
| `WebMCP.callTool` | Command | Invoke a tool by name with arguments. Routes to the corresponding [Antibot](https://scrapfly.io/docs/cloud-browser-api/cdp-reference/Antibot) handler. Returns `success`, `result`, and `errorMessage`. |
| `WebMCP.toolsAdded` | Event | Fired when tools become available. Each tool carries a `name`, `description`, and `inputSchema`. |
| `WebMCP.toolsRemoved` | Event | Fired when tools are removed. |
### Raw CDP Example
You can interact with the browser directly over CDP. Browser interaction tools use the `Antibot.*` CDP commands, while page-registered tools use `WebMCP.invokeTool`:
```
// 1. Call Antibot.clickOn directly via CDP:
{"id": 1, "method": "Antibot.clickOn", "params": {
"selector": {"type": "axNodeId", "query": "42"}
}}
// Result:
{"id": 1, "result": {"success": true}}
// 2. Enable WebMCP to discover page-registered tools:
{"id": 2, "method": "WebMCP.enable"}
// If the page registered tools, you receive toolsAdded:
{
"method": "WebMCP.toolsAdded",
"params": {
"tools": [
{"name": "addToCart", "description": "Add product to cart", "inputSchema": {...}, "frameId": "ABC123"},
{"name": "searchProducts", "description": "Search by keyword", "inputSchema": {...}, "frameId": "ABC123"}
]
}
}
// 3. Invoke a page-registered tool:
{"id": 3, "method": "WebMCP.invokeTool", "params": {
"frameId": "ABC123",
"toolName": "addToCart",
"input": {"title": "Widget Alpha"}
}}
// Response (invocationId):
{"id": 3, "result": {"invocationId": "inv-001"}}
// Then toolResponded event:
{"method": "WebMCP.toolResponded", "params": {
"invocationId": "inv-001",
"status": "Success",
"output": {"added": true, "product": "Widget Alpha"}
}}
```
### Available Tools
When a browser session is opened, the MCP server registers these tool categories:
#### Browser Interaction Tools (`action_*`)
Human-like browser interactions using the [Antibot CDP domain](https://scrapfly.io/docs/cloud-browser-api/cdp-reference/Antibot). All actions use realistic mouse movement (Bezier curves) and keyboard timing.
| Tool | Description |
|---|---|
| `clickOn` | Click on an element with human-like mouse movement |
| `fill` | Click a form field, optionally clear, then type text |
| `typeText` | Type text at the current cursor position |
| `pressKey` | Press a keyboard key (Enter, Tab, Ctrl+a, etc.) |
| `scroll` | Scroll element into view, to bottom, or by delta |
| `hover` | Hover over an element |
| `selectOption` | Select from a dropdown (native or custom) |
| `waitForElement` | Wait for an element to appear |
| See all actions in the [Antibot CDP Reference](https://scrapfly.io/docs/cloud-browser-api/cdp-reference/Antibot) |
#### Page-Registered WebMCP Tools (`webmcp_*`)
If the web page registers tools via `navigator.modelContext.registerTool()`, they are discovered automatically and exposed as `webmcp_*` tools. These are semantic page-level actions, for example, an e-commerce page might expose `addToCart`, `searchProducts`, or `getProductDetails`.
#### Native CDP Tools (`browser_*`)
| Tool | Description |
|---|---|
| `take_screenshot` | Capture the page as PNG (full page or element) |
| `evaluate_script` | Execute JavaScript and return the result |
| `take_snapshot` | Get page text content (DOM snapshot) |
| `get_performance_metrics` | Page load timing, resource counts, JS heap |
### Try It: Register and Call a Custom Tool
This end-to-end example connects to a Cloud Browser with MCP enabled, injects a custom tool into the page via `navigator.modelContext`, then discovers and calls it through the WebMCP CDP domain.
Python (Playwright) JavaScript (Puppeteer)
```
import asyncio
import json
from playwright.async_api import async_playwright
WS_URL = 'wss://browser.scrapfly.io?api_key=&enable_mcp=true'
async def main():
async with async_playwright() as pw:
browser = await pw.chromium.connect_over_cdp(WS_URL)
page = browser.contexts[0].pages[0]
# Navigate to any page
await page.goto('https://web-scraping.dev/mcp-tools')
# Register a custom MCP tool on the page
await page.evaluate('''() => {
navigator.modelContext.registerTool({
name: 'getProductCount',
description: 'Returns the number of products on the page',
inputSchema: { type: 'object', properties: {} },
handler: async () => {
const items = document.querySelectorAll('.product-item');
return { count: items.length };
}
});
}''')
# Get a CDP session
cdp = await page.context.new_cdp_session(page)
# Enable the WebMCP domain and collect tools
tools = []
cdp.on('WebMCP.toolsAdded', lambda params: tools.extend(params['tools']))
await cdp.send('WebMCP.enable')
# Print discovered tools
print(f'Discovered {len(tools)} tools:')
for tool in tools:
print(f' - {tool["name"]}: {tool["description"]}')
# Find and call our custom tool
result = await cdp.send('WebMCP.callTool', {
'name': 'getProductCount',
'arguments': '{}'
})
print(f'Result: {json.dumps(result)}')
await browser.close()
asyncio.run(main())
```
```
const puppeteer = require('puppeteer-core');
const WS_URL = 'wss://browser.scrapfly.io?api_key=&enable_mcp=true';
(async () => {
const browser = await puppeteer.connect({ browserWSEndpoint: WS_URL });
const [page] = await browser.pages();
// Navigate to any page
await page.goto('https://web-scraping.dev/mcp-tools');
// Register a custom MCP tool on the page
await page.evaluate(() => {
navigator.modelContext.registerTool({
name: 'getProductCount',
description: 'Returns the number of products on the page',
inputSchema: { type: 'object', properties: {} },
handler: async () => {
const items = document.querySelectorAll('.product-item');
return { count: items.length };
}
});
});
// Get a CDP session
const cdp = await page.createCDPSession();
// Enable the WebMCP domain and collect tools
const tools = [];
cdp.on('WebMCP.toolsAdded', (params) => tools.push(...params.tools));
await cdp.send('WebMCP.enable');
// Print discovered tools
console.log(`Discovered ${tools.length} tools:`);
tools.forEach(t => console.log(` - ${t.name}: ${t.description}`));
// Find and call our custom tool
const result = await cdp.send('WebMCP.callTool', {
name: 'getProductCount',
arguments: '{}'
});
console.log('Result:', JSON.stringify(result));
await browser.close();
})();
```
Expected output:
```
Discovered 18 tools:
- clickOn: Click on an element
- fill: Fill a form field
- type: Type text into an element
...
- getProductCount: Returns the number of products on the page
Result: {"success": true, "result": "{"count":5}", "errorMessage": null}
```
The output shows both the built-in Antibot tools and your custom `getProductCount` tool. The custom tool runs in the page context and returns structured data.
### Test with an AI Agent
The best way to validate your MCP setup is to let an AI agent discover and call tools autonomously. Using the [Scrapfly MCP Server](https://scrapfly.io/docs/mcp/getting-started), an agent like Claude can open a Cloud Browser session, discover the available WebMCP tools, and use them to interact with the page.
Try this prompt in the [MCP Playground](https://scrapfly.io/dashboard/playground/mcp) or with any MCP-compatible AI agent:
```
Open a cloud browser on https://web-scraping.dev/mcp-tools,
list the available page MCP tools,
and use them to search for products
```
The agent will:
1. Call `cloud_browser_open` with `enable_mcp=true`
2. Receive a `notifications/tools/list_changed` with the discovered WebMCP tools
3. See tools like `webmcp_{session}_clickOn`, `webmcp_{session}_fill`, etc. in `tools/list`
4. Call the tools to interact with the page (search, click, scroll)
5. Return the results as structured data
This validates the full chain: browser allocation, MCP endpoint, tool discovery, and tool invocation. If the agent can call a `webmcp_*` tool and get a result, your setup works.
**Tip:** Most users should connect to the [streamable-HTTP MCP endpoint](#mcp-endpoint) or use the [Scrapfly MCP Server](#mcp-server-integration) instead of using the CDP domain directly. The raw CDP domain is available for advanced use cases or when you need fine-grained control.
For the complete type definitions, parameters, and event payloads, see the [WebMCP CDP Reference](https://scrapfly.io/docs/cloud-browser-api/cdp-reference/WebMCP).
## Enabling MCP
Add `enable_mcp=true` to your Cloud Browser WebSocket URL:
```
wss://browser.scrapfly.io?api_key=&enable_mcp=true
```
Or use the SDK:
Python TypeScript Go Rust
```
from scrapfly import ScrapflyClient, BrowserConfig
client = ScrapflyClient(key='')
config = BrowserConfig(
enable_mcp=True,
)
ws_url = client.cloud_browser(config)
# ws_url contains the CDP WebSocket URL
# The MCP endpoint is returned in the allocation response
```
```
import { ScrapflyClient, BrowserConfig } from 'scrapfly-sdk';
const client = new ScrapflyClient({ key: '' });
const wsUrl = client.cloudBrowser({
enable_mcp: true,
});
// wsUrl contains the CDP WebSocket URL
// The MCP endpoint is returned in the allocation response
```
```
client, _ := scrapfly.New("")
wsURL := client.CloudBrowser(&scrapfly.CloudBrowserConfig{
EnableMCP: true,
})
// wsURL contains the CDP WebSocket URL
// The MCP endpoint is returned in the allocation response
```
```
use scrapfly::ScrapflyClient;
use scrapfly::cloud_browser::BrowserConfig;
let client = ScrapflyClient::new("");
let config = BrowserConfig {
enable_mcp: true,
..Default::default()
};
let ws_url = client.cloud_browser_url(&config);
// ws_url contains the CDP WebSocket URL
// The MCP endpoint is returned in the allocation response
```
## MCP Endpoint
When `enable_mcp=true` is set, the browser allocation response includes an `mcp_endpoint` field, a streamable-HTTP URL your AI agent can connect to:
```
{
"browser_version": "Chrome/147.0.7727.55",
"web_socket_debugger_url": "wss://...",
"mcp_endpoint": "https://...:9223/mcp",
"allocation_id": "..."
}
```
The MCP endpoint supports the [streamable-HTTP transport](https://modelcontextprotocol.io/specification/2025-03-26/basic/transports#streamable-http), which is the standard MCP transport for server-sent events over HTTP. Any MCP-compatible client can connect to it.
## Use Cases
- **AI agent browser automation** - let Claude, GPT, or custom agents interact with websites through structured tool calls instead of fragile selectors
- **Structured data extraction** - when a website exposes WebMCP tools, agents can call `getProducts()` or `search(query)` directly instead of scraping HTML
- **Form filling and checkout flows** - WebMCP tools provide typed schemas for form inputs, reducing errors in multi-step workflows
- **Testing AI-ready websites** - validate that your site's WebMCP tool registrations work correctly when called by an AI agent
## Compatibility
WebMCP is an early-preview feature available since Chrome 146 Canary. Scrapium ships with these feature flags enabled:
| Chrome Feature Flag | Purpose |
|---|---|
| `DevToolsWebMCPSupport` | Enables the WebMCP CDP domain in DevTools, allowing CDP clients to discover and invoke page-registered MCP tools |
| `WebMCPTesting` | Activates the `navigator.modelContext` JavaScript API for tool registration on all pages (not just flagged origins) |
Scrapium's MCP implementation is fully compliant with [chrome-devtools-mcp](https://github.com/ChromeDevTools/chrome-devtools-mcp). Any MCP client that works with chrome-devtools-mcp works with Scrapium out of the box.
**Early Preview:** WebMCP is a proposed standard under active development. The API surface and tool discovery behavior may change as the specification evolves through the W3C standardization process. See the [Chrome blog post](https://developer.chrome.com/blog/webmcp-epp) for the latest status.
## Scrapfly MCP Server Integration
If you use the [Scrapfly MCP Server](https://scrapfly.io/docs/mcp/getting-started), the browser's WebMCP tools are **automatically proxied** through your existing MCP connection. You don't need to connect to the Scrapium MCP endpoint directly.
#### How it works
1. **Open a browser session**Call [`cloud_browser_open`](https://scrapfly.io/docs/mcp/tools#cloud_browser_open) via the Scrapfly MCP Server. The server allocates a Cloud Browser and connects via CDP WebSocket.
2. **Browser interaction tools registered**Action tools (`action_*`) are registered statically, they call `Antibot.*` CDP commands directly with realistic mouse/keyboard timing.
3. **Native CDP tools registered**Native tools (`browser_*`) are registered for screenshot, evaluate JavaScript, accessibility snapshot, and performance metrics.
4. **Page WebMCP tools discovered**If the page registers tools via `navigator.modelContext.registerTool()`, they are discovered via `WebMCP.toolsAdded` events and registered as `webmcp_*`.
5. **Agent notified**Your agent receives a `notifications/tools/list_changed` notification with all available tools.
6. **Tool calls proxied**`action_*` → `Antibot.*` CDP, `webmcp_*` → `WebMCP.invokeTool` CDP, all through a single MCP connection.
```
AI Agent (Claude, GPT, ...)
│
│ Single MCP connection
▼
Scrapfly MCP Server
│
│ tools/list includes:
│ web_scrape, screenshot, ...
│ + webmcp_abc123_search ← from Scrapium
│ + webmcp_abc123_addToCart ← from Scrapium
│
│ tools/call proxied to Scrapium
▼
Scrapium (WebMCP)
│
│ Page executes tool
▼
Structured result
```
**Single connection, all tools:** Your agent connects to the Scrapfly MCP Server once and gets scraping tools, screenshot tools, AND page-specific WebMCP tools, all through the same MCP transport. No need to manage a second connection to Scrapium's MCP endpoint.
When you call `cloud_browser_navigate`, old tools are removed and new ones are discovered on the target page. When you call `cloud_browser_close`, all dynamic tools are unregistered.
[See the full Cloud Browser MCP tools documentation →](https://scrapfly.io/docs/mcp/tools#cloud_browser_open)
## Default Behavior
MCP is **disabled by default**. When `enable_mcp` is not set (or set to `false`), the browser launches without the WebMCP feature flags and no MCP endpoint is exposed. This ensures zero overhead for sessions that don't need AI agent integration.
## Connection Parameters
| Parameter | Required | Default | Description |
|---|---|---|---|
| `enable_mcp` | No | `false` | Enable Scrapium's built-in MCP support. When `true`, the browser is launched with WebMCP feature flags and an MCP streamable-HTTP endpoint is exposed in the allocation response. |
See the full list of connection parameters in the [Getting Started](https://scrapfly.io/docs/cloud-browser-api/getting-started#connection-parameters) guide.