home / mcp / playwright scraper mcp server
Provides an MCP server to scrape web pages and convert content to Markdown using Playwright, BeautifulSoup, and pypandoc.
Configuration
View docs{
"mcpServers": {
"dennisgl-mcp-playwright-scraper": {
"command": "uvx",
"args": [
"mcp-playwright-scraper"
]
}
}
}You run an MCP server that can crawl web pages and convert the content into Markdown, enabling you to capture clean, readable output from live sites while handling modern JavaScript-heavy pages.
Once you have the Playwright-based MCP server configured, you can prompt your MCP client to use it for scraping and Markdown conversion. Request content from a URL and have the server return the result as Markdown. You can combine this with summaries or analysis as part of your workflow by instructing the server to process the page and then deliver the formatted output.
Prerequisites you need before installing the MCP server:
- Python 3.11 or higher.
- Playwright browser dependencies (you will install chromium support).
- Pandoc support is optional; pypandoc will install it automatically if possible.
# Install the MCP server package that provides the scraping tool
pip install mcp-playwright-scraper
# Run the server as a module
python -m mcp_playwright_scraper
# Install Playwright browser dependencies (Chromium)
playwright install --with-deps chromiumYou configure how the MCP server runs in your environment by using a client-specific launcher. The server exposes a stdio-based interface that you can start locally.
If you use the provided CLI helpers, you can start the server with one of these commands:
{
"type": "stdio",
"name": "playwright_scraper",
"command": "uvx",
"args": ["mcp-playwright-scraper"]
}You can run the server using different entry points depending on your setup. For a standard uvx-based workflow, start the server with the directory option and the run command. If you prefer the Python module approach, start with the Python module invocation.
# uvx-based local run (example)
uvx mcp-playwright-scraper
# Alternatively, using a directory layout (development/unpublished example)
uv run mcp-playwright-scraper --directory /path/to/mcp-playwright-scraper
# Python module run (pip installation example)
python -m mcp_playwright_scraperScrapes content from a URL and converts it to Markdown. Optional parameter verify_ssl controls SSL certificate verification (default: true).