- Home
- MCP servers
- Cloudflare Playwright MCP Example
Cloudflare Playwright MCP Example
- typescript
0
GitHub Stars
typescript
Language
7 months ago
First Indexed
3 months ago
Catalog Refreshed
Documentation & install
Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.
You can run a Cloudflare-based Playwright browser controller as an MCP server so AI assistants can navigate web pages, type, click, and capture screenshots. This server exposes a remote HTTP endpoint for MCP clients and can be proxied locally for desktop tools, enabling seamless browser automation across platforms.
How to use
Connect to the server from your MCP client using the remote URL path exposed by the worker. This enables you to run browser automation tasks through standard MCP tool calls, such as navigating to pages, typing text, clicking elements, and taking screenshots.
You can also integrate with desktop tooling by proxying the remote MCP server locally. For Claude Desktop, you can route requests through a local proxy so the desktop app can communicate with this MCP server.
How to install
Prerequisites: ensure you have Node.js and npm installed on your machine. You also need the Cloudflare Wrangler tool to deploy Workers.
Install dependencies and deploy the MCP server to Cloudflare Workers.
# Install dependencies
npm ci
# Deploy to Cloudflare Workers
npx wrangler deploy
MCP connection configurations
Configure how you connect to this MCP server in your client or tooling. You have two explicit connection methods: a remote HTTP endpoint and a local stdio proxy for desktop tooling.
HTTP (remote) endpoint example URL pattern you will connect to from MCP clients.
MCP connections (explicit configurations)
{
"type": "http",
"name": "cloudflare_playwright_mcp",
"url": "https://[my-mcp-url].workers.dev/sse",
"args": []
}
{
"type": "stdio",
"name": "cloudflare_playwright_mcp",
"command": "npx",
"args": ["mcp-remote", "https://[my-mcp-url].workers.dev/sse"]
}
Available tools
browser_navigate
Navigate the active browser to a specified URL and render the resulting page.
browser_type
Type text into the focused element, supporting simple keystrokes and input.
browser_clickt
Click a target element, such as a button or checkbox, identified by its label or selector.
browser_take_screenshot
Capture a screenshot of the current browser view and return or save it for review.