mcpField
M
MCP Server MetaTool

MCP Server MetaTool

Local
github.com/metatool-aiDeveloper Tools5Updated 87w ago
Repository

About MCP Server MetaTool

MCP Server MetaTool is an unmaintained proxy that combines tools from multiple MCP servers and routes calls to the corresponding server. It gets server configuration from metatool-app and is intended to be used with that GUI manager.

Key features

  1. 01Combines tools from multiple servers
  2. 02Routes calls to the matching server
  3. 03Retrieves configuration from metatool-app
  4. 04Separates tools by namespace

Use cases

  1. 01Presenting several MCP servers to one client
  2. 02Routing tool calls across configured servers

MCP support

Acts as both an MCP client and an MCP server: it connects to configured MCP servers, presents their tools to MCP clients, and forwards calls. The project is no longer maintained.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
  "mcpServers": {
    "metatool": {
      "command": "uvx",
      "args": [
        "mcp-server-metatool"
      ]
    }
  }
}
Run commanduvx mcp-server-metatool

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
pypimcp-server-metatool

Tools

No tool list yet

Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.

Details

From the project README

MCP Server MetaTool

Warning
⚠️ This repo is no longer maintained!

Check out this Node.js substitute: metatool-ai/mcp-server-metamcp.
It has way better performance and fewer bugs.
(I am the author there too.)

Overview

Metatool MCP Server is a proxy server that joins multiple MCP servers and forward tool calls to the appropriate server. It should be used with metatool-app, the GUI tool manager for MCP which is also open source, together.

sequenceDiagram
    participant MCPClient as MCP Client (e.g. Claude Desktop)
    participant MetaToolMCP as MetaTool MCP Server
    participant MetaToolApp as MetaTool App
    participant MCPServers as Installed MCP Servers in Metatool App

    MCPClient ->> MetaToolMCP: Request list tools
    MetaToolMCP ->> MetaToolApp: Get tools configuration & status
    MetaToolApp ->> MetaToolMCP: Return tools configuration & status

    loop For each listed MCP Server
        MetaToolMCP ->> MCPServers: Request list_tools
        MCPServers ->> MetaToolMCP: Return list of tools
    end

    MetaToolMCP ->> MetaToolMCP: Aggregate tool lists
    MetaToolMCP ->> MCPClient: Return aggregated list of tools

    MCPClient ->> MetaToolMCP: Call tool
    MetaToolMCP ->> MCPServers: call_tool to target MCP Server
    MCPServers ->> MetaToolMCP: Return tool response
    MetaToolMCP ->> MCPClient: Return tool response

How it works

  • mcp-server-metatool itself is both a MCP client and a MCP server
  • On startup, it connects to the metatool-app API to get a list of MCP servers to connect to
  • On list_tools call of MCP, it joins all tools from all MCP servers with namespace isolation.
  • On call_tool call of MCP, it forwards the call to the appropriate server.

Installation

Best way to use this is through uv A METATOOL_API_KEY environment variable must be set

export METATOOL_API_KEY="my_api_key" # get from metatool-ai/metatool-app, which is self hostable and free
uvx mcp-server-metatool

You need a MCP Client to connect to this server to use it.

Additional Configuration

A METATOOL_API_BASE_URL environment variable can be set to point to another metatool instance

export METATOOL_API_BASE_URL="http://localhost:12005"

License

Apache License 2.0

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.