Product · MCP Tools

Fetchium as an MCP server for AI clients

The Model Context Protocol (MCP) lets AI clients like Claude Desktop and Cursor call external tools without writing any integration code. Fetchium ships a JSON-RPC 2.0 stdio MCP server with 12 tools spanning search, fetch, estimate, research, YouTube, and social workflows.

Get API Key Free →MCP Setup Guide

12 MCP Tools

Fetchium exposes 12 tool definitions in the current MCP server. A representative subset is shown below.

fetchium_search

Search the web via 17 backends. Returns ranked results with content snippets.

query: stringmax_results?: numberbackends?: string[]

fetchium_fetch

Fetch and extract clean content from a URL using the CEP pipeline.

url: stringquery?: stringtoken_budget?: number

fetchium_estimate

Estimate token and processing costs before a request.

query?: stringurl?: stringtoken_budget?: number

fetchium_research

Run a full multi-agent research task and return a structured report with citations.

query: stringdepth?: 'standard' | 'deep'max_sources?: number

fetchium_youtube_search

Search YouTube content and return structured video results.

query: stringmax_results?: number

fetchium_youtube_analyze

Analyze a YouTube video, transcript, or channel context.

url?: stringvideo_id?: string

fetchium_social_research

Run cross-platform social research workflows.

query: stringmax_results?: number

Setup: Claude Desktop

claude_desktop_config.json

{
  "mcpServers": {
    "fetchium": {
      "command": "fetchium",
      "args": ["mcp"],
      "env": {
        "FETCHIUM_API_KEY": "your_api_key_here"
      }
    }
  }
}

Install the CLI: npm install -g fetchium or cargo install fetchium