Metaso

You run a Metaso MCP Server to empower large language models with AI search, webpage reading, and AI-assisted Q&A. This server focuses on fast, secure access to Metaso’s search capabilities and enables seamless integration with your MCP clients for richer, context-aware responses.

How to use

You connect to the MCP server from your client by configuring a local or remote MCP endpoint. You will typically provide an API key, and then issue search, read, or chat requests through your MCP client. The server exposes three core capabilities: AI search across diverse content ranges, webpage content extraction, and AI-assisted Q&A built on search results. Use the search tool to gather sources, the reader to fetch complete page content, and the chat tool to get contextual answers that reference those sources. Ensure your client passes the required API key for authentication and respects content length and safety checks.

How to install

Prerequisites you need before installing: Node.js and npm, Git, and a suitable shell environment. You will run commands to clone the project, install dependencies, and build the server before starting it.

# Clone the MCP server repository
git clone https://github.com/csrts/metaso-mcp
cd metaso-mcp

# Install dependencies
npm install

# Build the project
npm run build

Configuration and startup

Set up your environment with the required API key and optional settings to customize timeouts and debug logs. The key controls access to the Metaso API, and you can enable verbose debugging during development.

export METASO_API_KEY="mk-YOUR_API_KEY_HERE"
export METASO_BASE_URL="https://metaso.cn"
export METASO_TIMEOUT=30000
export METASO_DEBUG=false

Start the MCP server in STDIO mode

Choose a development or production runtime. The server can run as a local process that your MCP client starts and communicates with through standard I/O.

# Development/studio mode example (local TS/JS file)
"mcpServers": {
  "metaso": {
    "command": "node",
    "args": ["/path/to/metaso-mcp-server/dist/index.js"],
    "env": {
      "METASO_API_KEY": "mk-YOUR_API_KEY_HERE"
    }
  }
}

Available tools and how to use them

The MCP server provides three core tools you can invoke from your client: AI search, webpage reading, and AI chat.

Available tools

metaso_search

AI search across webpages, documents, scholars, images, videos, and podcasts with options for page-based or size-based results and optional summaries.

metaso_reader

Fetch complete content of a specified webpage and output in markdown or json format.

metaso_chat

AI-assisted Q&A that leverages search results to provide context-rich answers; supports multiple models and scopes.