SearXNG

You can use this MCP server to enable AI assistants to perform web searches through a SearXNG instance. It exposes a simple API to run searches, discover available search engines, and configure search parameters like categories, languages, and safe search. This makes your AI workflows more capable by leveraging your existing SearXNG deployment.

How to use

You interact with the server using an MCP client. Point the client at the MCP endpoint you run (either via an HTTP URL or a local stdio integration) and send search requests through the standardized MCP methods. You can perform a web search, retrieve the list of available engines, and adjust optional search parameters such as categories, languages, and safe search. When a client calls the search function, it returns results from your SearXNG instance; when calling the engine discovery function, you can adapt your UI to present the available engines to users.

Get started with a remote MCP endpoint

• Choose how you want to connect: HTTP (remote MCP endpoint) or stdio (local MCP runtime).
• Use an MCP client to query searches or list engines.
• If you’re behind a firewall or want centralized access, prefer the HTTP method.

Available tools

search

Performs a web search against the configured SearXNG instance. You can specify optional categories, engines, language, page, time range, and safe search settings to filter results.

get_available_engines

Retrieves the list of search engines available in your SearXNG instance so you can present them to users or tailor UI options.