Skip to main content
The Model Context Protocol (MCP) is an open standard that lets AI agents and assistants connect to external tools and data sources through a uniform interface. fetchbean exposes an MCP server so you can plug it directly into any MCP-compatible AI agent or assistant — Claude, a custom LangChain agent, or any other framework that supports MCP. Instead of configuring separate integrations for web search, page reading, weather, and more, you point your agent at the fetchbean MCP server and it handles everything through a single connection.

​
What you get

The fetchbean MCP server provides two tools your agent can call:

search_fetchbean

Semantic search across the fetchbean knowledge base. Use this to find relevant documentation, code examples, API references, and guides. Returns contextual content with titles and direct links to source pages.

query_docs_filesystem_fetchbean

Read-only, shell-like access to the fetchbean documentation filesystem. Supports rg, grep, find, tree, ls, cat, head, tail, and other standard text utilities against a virtualized docs filesystem. Each call is stateless — the working directory resets to / on every invocation.

​
Server configuration

Add the fetchbean MCP server to your agent’s configuration. The server communicates over HTTP transport so no local process is required. 
{
  "server": {
    "name": "fetchbean",
    "version": "1.0.0",
    "transport": "http"
  }
}
Once connected, your agent can call search_fetchbean for broad conceptual queries and query_docs_filesystem_fetchbean for exact keyword matching, structural exploration, or reading the full content of a specific page by path.

​
Tool reference

​
search_fetchbean

Search the fetchbean knowledge base to find relevant information, code examples, API references, and guides. Pass a natural-language query. If you need the full content of a specific page after searching, use query_docs_filesystem_fetchbean to head or cat it — append .mdx to the path returned from search results (for example, head -200 /guides/raw-run.mdx).
InputTypeRequiredDescription
querystringyesA natural-language query to search the knowledge base with.

​
query_docs_filesystem_fetchbean

Run a read-only shell-like command against a virtualized, in-memory filesystem rooted at / that contains only the fetchbean documentation pages and OpenAPI specs. This is not a shell on any real machine — it is a sandbox backed by documentation content. Use this tool to read documentation pages by path, search with exact keyword or regex matches using rg, or explore the docs structure with tree or ls. Pass the .mdx path to head or cat to read a page (for example, head -80 /quickstart.mdx).
InputTypeRequiredDescription
commandstringyesA shell command to run against the docs filesystem, e.g. head -80 /guides/raw-run.mdx or rg -il "rate limit" /.
Supported commands: rg (ripgrep), grep, find, tree, ls, cat, head, tail, stat, wc, sort, uniq, cut, sed, awk, jq, and basic text utilities. No writes, no network access.

​
When to use the MCP server

The fetchbean MCP server is the right choice when your agent:
  • Needs real-time data such as web search results, page content, or weather.
  • Wants to avoid managing multiple API keys and provider SDKs.
  • Benefits from unified billing and rate limiting across all providers.
  • Runs inside an MCP-compatible host such as Claude, a custom LangChain agent, or any framework that supports the Model Context Protocol.
The MCP server is designed for read and retrieval workflows. For write operations or providers that require BYOK connections (Fireflies, PostHog, Linear), call POST /v1/run directly using your fetchbean API key. See the Raw Run guide.

​
Workflow tips

When using the MCP server in an agent, start with search_fetchbean for broad or conceptual queries like “how to authenticate” or “rate limiting”. Switch to query_docs_filesystem_fetchbean when you need exact keyword or regex matching, want to browse the documentation structure with tree, or need to read a full page — pass the .mdx path to head or cat.
You can access the MCP server configuration through the contextual menu on any fetchbean docs page. Look for the copy, view, ChatGPT, Claude, and MCP options to get the connection details for your specific agent setup.