Broken Link Checker

This MCP server scans web pages and domains to identify broken links, providing structured results that help you keep sites healthy and navigable. You can check a single page or crawl an entire domain, with output that includes the status of each link and where it was found.

How to use

You use this MCP server by connecting a client that supports MCP to the HTTP endpoint. Start by running the server, then issue commands to check either a single page or an entire domain. When you check a page, you’ll get a report listing every link on that page with its status and location. When you check a domain, the server crawls pages under that domain and compiles a comprehensive set of results. Use the two available tools to perform your checks: check_page for a single URL and check_domain for recursive domain-wide checks.

How to install

Prerequisites: you need Python 3 and pip installed on your system.

Install the MCP server package in editable mode to set up a local development environment.

pip install -e .

Start the MCP server using the Python entry point. The server will run on http://127.0.0.1:8000 by default.

python src/server.py

Additional sections

Output from the checks is structured to show the page URL, the link reference text, the link URL, and the status (for example, Bad or Good). A summary aggregates totals such as total links, good links, bad links, and pages scanned.

Robots.txt rules are respected during crawling, so you won’t probe pages that are disallowed by a site’s robots.txt file. The server detects hyperlinks, images, scripts, stylesheets, and media, ensuring you get a complete view of potential broken references.

If you need to run multiple checks or integrate into a larger workflow, you can reuse the same server endpoint to perform repeated validations across pages and domains. The tooling supports clear, structured output that your pipelines can parse and report on.

Available tools

check_page

Check all links on a single page by providing the page URL. Returns per-link status and location.

check_domain

Recursively check all pages within a domain starting from a root URL. Supports a crawl depth parameter to limit traversal.