playwright-mcp-fetch

Current version: 0.1.5

This tool provides a Model Context Protocol (MCP) server for fetching content from websites and converting it to different formats using Playwright.

Requirements

• Python 3.10 or higher

Features

• fetch_html: Fetch the raw HTML content from a website
• fetch_markdown: Fetch content from a website and convert it to Markdown format
• fetch_txt: Fetch and return plain text content from a website (HTML tags removed)
• fetch_json: Fetch and parse JSON content

Installation

From Source

git clone https://github.com/kevinwatt/playwright-mcp-fetch.git
cd playwright-mcp-fetch
pip install -e .

Install Dependencies

pip install -r requirements.txt
# Install Playwright browsers
playwright install

Usage

Run as stdio MCP Server

playwright-mcp-fetch

Run as SSE MCP Server

# Run with the SSE transport
playwright-mcp-fetch-sse

This will start an HTTP server providing the following endpoints:

• GET /: Server status page
• GET /sse: SSE connection endpoint
• POST /api/list-tools: List available tools
• POST /api/call-tool: Call a tool

Environment Variables

• PORT: HTTP server port (default: 3000)
• TRANSPORT_TYPE: Transport type, either stdio or sse (default: stdio)
• fetch_html: Whether to enable the fetch_html tool, either Enable or Disable (default: Disable)

MCP Client Configuration

To use this server in an MCP client, use the following configuration:

SSE Transport

{
  "mcpServers": {
    "fetch-tools": {
      "enabled": true,
      "transport": "sse",
      "url": "http://localhost:3000/sse"
    }
  }
}

stdio Transport

{
  "mcpServers": {
    "fetch-tools": {
      "enabled": true,
      "transport": "stdio",
      "command": "uvx",
      "args": ["playwright-mcp-fetch"]
    }
  }
}

You can replace fetch-tools with any name you prefer, and adjust the URL or command according to your deployment environment.

API Examples

List Tools

curl -X POST http://localhost:3000/api/list-tools

Call a Tool

curl -X POST http://localhost:3000/api/call-tool \
  -H "Content-Type: application/json" \
  -d '{"name": "fetch_markdown", "arguments": {"url": "https://example.com"}}'

SSE Client Example

import json
import aiohttp
import asyncio

async def sse_client():
    async with aiohttp.ClientSession() as session:
        async with session.get("http://localhost:3000/sse") as response:
            async for line in response.content:
                if line.startswith(b"data: "):
                    data = json.loads(line[6:].decode("utf-8"))
                    print(f"Received event: {data}")

asyncio.run(sse_client())

Development

# Install in development mode
pip install -e .

# Run tests
pytest

Docker Support

Build and run with Docker:

docker build -t playwright-mcp-fetch .
docker run -p 3000:3000 -e TRANSPORT_TYPE=sse playwright-mcp-fetch

Or use Docker Compose:

docker-compose up -d

License

MIT