What it does
crawlport gives an agent access to the readable web through Firecrawl. A single page comes back as clean markdown. For a whole site it can map discovered pages, crawl a section, or scrape a list of URLs in batch. For a question it can run a web search, extract structured fields from pages into a schema the user describes, research across many sources, or draft the llms.txt file a site would publish about itself.
It suits work where web content needs to become text or structured data for later steps, without handling page fetching directly. It runs as a command-line tool or as a stdio MCP server exposing ten tools: scrape, map, search, crawl, crawl status, extract, batch scrape, batch status, deep research and llms.txt generation. Optional arguments are passed on only when set, so the API's defaults apply otherwise.
Setup requires a Firecrawl API key, exported as an environment variable. Crawl and batch jobs are asynchronous: the start call returns a job id, and a matching status call reads progress and results. Successful CLI calls print the API's JSON response, and tool results are the upstream JSON. The project is released under the MIT license.
Limitations and requirements
Requires a Firecrawl API key set in the FIRECRAWL_API_KEY environment variable; without it every command and tool call fails and nothing is sent. Crawl and batch jobs run asynchronously and must be polled with status calls to get results.
Safety
Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.
- Runs code or changes an assistant's behaviour
Sources
- Official MCP Registry first seen 13 Nov 2026 · last seen 13 Nov 2026 · its page on the source
io.github.CloudCorpRecords/crawlport
Firecrawl web data: scrape, map, search, crawl, extract, research, and llms.txt.
v0.1.0As shown on registry.modelcontextprotocol.io when we last read it