Geonode MCP Server

What Geonode MCP Server does

Geonode MCP Server gives MCP clients tools for finding and collecting web content. The extract tool turns a URL into HTML or Markdown, allowing an assistant to use a page as context. Authenticated clients can also search the web, discover links beneath a site, crawl pages breadth-first, and submit explicit URL lists for batch extraction.

The service supports both immediate and long-running workflows. Synchronous operations return their results directly, while asynchronous extraction, crawl, and batch jobs return an ID that can be checked later. Separate tools expose job status, paginated results, cancellation, and account-level usage statistics.

How it works

Connect an MCP client to https://mcp.geonode.com/mcp. Basic page extraction can be attempted without an API key, but anonymous requests are synchronous and subject to small per-client rate and daily limits. Authenticated requests send the key in the X-Api-Key header and use the limits and features associated with the Geonode plan.

Page extraction accepts optional browser rendering, proxy geo-targeting, custom HTTP headers, and browser wait controls. JavaScript rendering is useful for pages whose content is loaded in the browser. A request can either return content immediately or create an asynchronous extraction job for later polling.

Crawls begin at a seed URL and follow links breadth-first up to a requested depth and page limit. URL mapping combines sitemap parsing with link extraction from the seed page. Search returns ranked results with locale, safe-search, pagination, recency, and engine-selection options.

Setup and configuration

Add the MCP endpoint to a compatible client that supports remote MCP connections. For anonymous testing, no credential is needed for the limited extraction path. To use authenticated tools, sign up at Geonode and configure the resulting API key so the client sends it as X-Api-Key.

The endpoint is:

https://mcp.geonode.com/mcp

Authenticated features include search, mapping, crawling, batch extraction, asynchronous job management, statistics, JavaScript rendering, and broader request options. The key itself is transmitted as a request header rather than represented by a documented server environment variable.

Tools and capabilities

• extract: Return a page as HTML or Markdown, with optional rendering, headers, geo-targeting, and wait behavior.
• job and jobs: Retrieve one asynchronous extraction or list extraction jobs with filters and pagination.
• batch, batch_status, and cancel_batch: Process an explicit URL list asynchronously, inspect paginated results, or stop new work from being scheduled.
• crawl, crawl_status, and cancel_crawl: Crawl a site from a seed URL, inspect results, or cancel further scheduling.
• map, map_job, and map_jobs: Discover site URLs and inspect individual or listed mapping jobs.
• search, search_job, and search_jobs: Run ranked web searches and inspect search job records and results.
• statistics: View aggregated extraction counts, outcomes, and usage over a date range.

Limitations and notes

Anonymous extraction cannot use JavaScript rendering or arbitrary request headers beyond Accept-Language. Authenticated asynchronous operations require an API key. Mapping is scoped to the seed host and path by default; subdomains can be included explicitly. Search filtering by time_range limits result recency, while the search option in URL mapping filters discovered links by substring rather than querying a search engine.