Scraper MCP Installation Guide

Overview

Connect Geonode scraping tools to AI assistants and IDEs through Model Context Protocol (MCP).

Connect Geonode's scraping tools to the AI assistants and IDEs you already use. Once connected, your assistant can extract single pages, batch-process URL lists, and crawl whole sites from inside the tool you already work in.

What is MCP?

Model Context Protocol (MCP) is an open standard that lets AI assistants connect to external tools and data sources. Connecting Geonode over MCP lets your assistant fetch, render, and structure live web pages through Geonode's residential-proxy network.

What you can do with Geonode MCP

Once connected, your assistant can call these tools based on your request:

ToolWhat it does
extractScrape a single URL into Markdown or HTML. Supports JS rendering, proxy geo-targeting, and custom headers.
batchAsynchronous extraction of up to 1,000 URLs in one job.
batch_statusPoll a batch job for progress and paginated results.
cancel_batchStop a running batch job.
crawlCrawl a site from a seed URL (BFS, depth up to 10, up to 10,000 pages, same-domain or subdomain filtering).
crawl_statusPoll a crawl job for progress and crawled pages.
cancel_crawlStop a running crawl job.
jobRetrieve the result of a single async extraction.
jobsList your extraction jobs with filters (status, URL, date, format).
statisticsAggregated usage stats: extraction count, success rate, tokens used.

The assistant picks the tool for you: ask it to read a page and it scrapes; ask it to pull every doc page on a site and it crawls.

Install Geonode MCP

Start with Before You Start for your API key, endpoint, and authentication. Then follow the guide for your client:

If something does not connect, see Help & FAQ.

On this page