Cloudflare Browser Rendering MCP Server

What is Cloudflare Browser Rendering MCP Server?

This MCP (Model Context Protocol) server provides tools for fetching and processing web content using Cloudflare Browser Rendering for use as context in LLMs. It is designed to work with both Claude and Cline client environments.

How to use Cloudflare Browser Rendering MCP Server?

Install via Smithery (npx -y @smithery/cli install @amotivv/cloudflare-browser-rendering-mcp --client claude) or manually from GitHub. Configure the BROWSER_RENDERING_API environment variable with the URL of a deployed Cloudflare Worker that has Browser Rendering and KV bindings. Once configured, tools like fetch_page, search_documentation, extract_structured_content, summarize_content, and take_screenshot become available in the client.

Key features of Cloudflare Browser Rendering MCP Server

• Fetch and process web pages for LLM context
• Search Cloudflare documentation for relevant content
• Extract structured content using CSS selectors
• Summarize web content for concise LLM context
• Take screenshots of web pages
• Works with both Claude Desktop and Cline

Use cases of Cloudflare Browser Rendering MCP Server

• A developer asking an LLM to summarize the latest Cloudflare docs on Browser Rendering
• An AI assistant extracting the main heading and first paragraph from a web page
• A user requesting a screenshot of a page to reference offline
• A researcher fetching and condensing long articles into a few sentences
• A knowledge worker searching Cloudflare documentation for specific API details

FAQ from Cloudflare Browser Rendering MCP Server

What are the prerequisites?

Node.js v18 or higher, a Cloudflare account with Browser Rendering API access, and a Cloudflare Worker deployed with the provided puppeteer-worker.js file, plus a KV namespace named SCREENSHOTS.

How do I configure the BROWSER_RENDERING_API environment variable?

Set the environment variable in the MCP server configuration (e.g., claude_desktop_config.json or cline_mcp_settings.json) with the URL of your deployed Cloudflare Worker, like "BROWSER_RENDERING_API": "https://your-worker-url.workers.dev".

What tools are available?

Five tools: fetch_page, search_documentation, extract_structured_content, summarize_content, and take_screenshot. Each accepts specific parameters listed in the README.

What Cloudflare Worker bindings are required?

A Browser Rendering binding named browser and a KV namespace binding named SCREENSHOTS. Without these, the worker will return errors.

How can I test that everything is working?

Run npm test after building the project and setting the BROWSER_RENDERING_API environment variable. This starts the MCP server and tests each tool with sample requests.