AI Meta

A dynamic MCP server that allows AI to create and execute custom tools through a meta-function architecture

javascript

GitHub Stars

javascript

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "alxspiker-ai-meta-mcp-server": {
      "command": "npx",
      "args": [
        "-y",
        "ai-meta-mcp-server"
      ],
      "env": {
        "PERSIST_TOOLS": "true",
        "TOOLS_DB_PATH": "./tools.json",
        "ALLOW_JS_EXECUTION": "true",
        "ALLOW_SHELL_EXECUTION": "false",
        "ALLOW_PYTHON_EXECUTION": "false"
      }
    }
  }
}

You run an AI Meta MCP Server that lets AI models define and execute their own custom tools at runtime. Tools run in sandboxed environments, can be persisted between sessions, and require human approval for creation and execution, helping you safely extend AI capabilities within controlled boundaries.

How to use

You interact with the MCP server through a client that can manage, register, and execute tools defined by AI. Tools can be created, listed, updated, and deleted. The server supports multiple runtimes for running custom tools, and you control which runtimes are enabled via environment variables. When a tool is created or executed, a human in the loop must approve the action.

How to install

Prerequisites: you need Node.js and npm installed on your system. Optionally, you can run the server directly with npx or install via Docker for isolated environments.

# Install the MCP server package locally in your project
npm install ai-meta-mcp-server

# Run the server directly with npx (no local install required)
npx ai-meta-mcp-server

# Or install and run via Docker (see Docker-specific steps below)

Running with Docker

Use Docker to run the MCP server with isolated storage and controlled permissions.

# Build the Docker image
docker build -t ai-meta-mcp-server .

# Run the container
docker run --rm -i ai-meta-mcp-server

# Run with custom configuration and persistent storage
docker run --rm -i \
  -e ALLOW_PYTHON_EXECUTION=true \
  -e ALLOW_SHELL_EXECUTION=false \
  -v $(pwd)/data:/app/data \
  ai-meta-mcp-server

Running with Claude Desktop

If you use Claude Desktop, specify the MCP server command to start the server from Claude’s configuration.

{
  "mcpServers": {
    "ai_meta_mcp": {
      "command": "npx",
      "args": ["-y", "ai-meta-mcp-server"],
      "env": {
        "ALLOW_JS_EXECUTION": "true",
        "ALLOW_PYTHON_EXECUTION": "false",
        "ALLOW_SHELL_EXECUTION": "false"
      }
    }
  }
}

Available tools

define_function

Meta-tool used to create a new tool by defining its function and behavior, enabling dynamic tool registration at runtime.

calculate_compound_interest

Tool that computes compound interest given principal, rate, time, and compounding frequency.

Built by

VeilStrat

AI signals for GTM teams