deno2-playwright MCP Server

Use This MCP server To

Automate web page interactions for AI agents Capture screenshots of web content programmatically Execute JavaScript on live web pages via LLMs Enable real-time web scraping with browser context Integrate browser automation in AI-enhanced workflows Test web applications through scripted browser control Run cross-platform browser automation binaries Support multi-step reasoning with live web data

README

Deno 2 Playwright Model Context Protocol Server Example

A Model Context Protocol server that provides browser automation capabilities using Playwright. This server enables LLMs to interact with web pages, take screenshots, and execute JavaScript in a real browser environment.

This repo uses Deno 2, which has nice ergonomics, because you can compile a binary and run it without any runtime dependencies.

This code is heavily based on the official Puppeteer MCP server, which you can find here: https://github.com/modelcontextprotocol/servers/tree/main/src/puppeteer

How to build

Only the mac binary build has been tested, but you should be able to build an executable binary for linux x86_64, linux ARM64, and windows x86_64.

deno task build-mac
deno task build-linux-x86_64
deno task build-linux-ARM64
deno task build-windows-x86_64

How to run

To invoke the playwright-server binary, you need to update your ~/Library/Application\ Support/Claude/claude_desktop_config.json to point to the binary.

{
  "mcpServers": {
    "playwright": {
      "command": "/path/to/deno2-playwright-mcp-server/playwright-server"
    }
  }
}

deno2-playwright-mcp-server FAQ

How do I build the deno2-playwright-mcp-server binary?

Use the provided Deno tasks like `deno task build-mac` for macOS or equivalent tasks for Linux and Windows to compile platform-specific binaries.

Does the server require any runtime dependencies?

No, the server is built with Deno 2, allowing you to compile a standalone binary with no runtime dependencies.

Can this server run on multiple operating systems?

Yes, it supports macOS, Linux (x86_64 and ARM64), and Windows (x86_64) platforms.

How does this server enable LLMs to interact with web pages?

It exposes browser automation capabilities via Playwright, allowing LLMs to navigate pages, execute JavaScript, and take screenshots.

Is this server based on any existing MCP server implementations?

Yes, it is heavily based on the official Puppeteer MCP server implementation.

How do I run the server after building it?

After building, update your client configuration (e.g., Claude desktop config) to point to the playwright-server binary and start it accordingly.

What advantages does using Deno 2 provide for this MCP server?

Deno 2 offers ergonomic development, easy binary compilation, and no external runtime dependencies, simplifying deployment.

Can this server execute arbitrary JavaScript on web pages?

Yes, it allows execution of JavaScript within the browser context to enable dynamic interactions.

Is the source code publicly available?

Yes, the source is hosted on GitHub, allowing inspection and customization.

How does this server improve AI workflows?

By enabling real-time, interactive web automation, it enhances AI agents' ability to gather and manipulate live web data.