AI Meta MCP Server

What is AI Meta MCP Server?

AI Meta MCP Server is a dynamic MCP (Model Context Protocol) server that allows AI models to create and execute their own custom tools at runtime. It provides a meta-function architecture enabling AI to extend its own capabilities by defining custom functions. This server is designed for developers and AI systems that need flexible, on-the-fly tool creation within a secure environment.

How to use AI Meta MCP Server?

Install via npm: npm install ai-meta-mcp-server. Run with npx ai-meta-mcp-server or with Docker. Configure behavior using environment variables (e.g., ALLOW_JS_EXECUTION, ALLOW_PYTHON_EXECUTION, PERSIST_TOOLS). For Claude Desktop, add the server to claude_desktop_config.json with command npx -y ai-meta-mcp-server and desired env settings.

Key features of AI Meta MCP Server

• Dynamic tool creation by AI models at runtime
• Multiple runtime environments: JavaScript, Python, Shell
• Sandboxed security for tool execution
• Persistent storage of custom tool definitions
• Flexible tool registry for managing, listing, updating, deleting tools
• Human approval flow for tool creation and execution

Use cases of AI Meta MCP Server

• AI assistants creating custom calculation tools on demand
• Extending AI capabilities with user-defined functions at runtime
• Running code in isolated sandboxes for safe experimentation
• Persisting custom tools across sessions for repeated use

FAQ from AI Meta MCP Server

What runtime environments are supported for tool execution?

JavaScript, Python, and Shell execution are supported, configurable via environment variables (default: JS enabled, Python and Shell disabled).

How are tools secured?

All code executes in sandboxed environments. Human approval is required for tool creation and execution. Tool execution privileges are configurable via environment variables, and audit logging tracks all operations.

Can tools be persisted between sessions?

Yes, by default tools are saved to a JSON database (TOOLS_DB_PATH). Set PERSIST_TOOLS=true to enable persistence.

How do I run the server?

Install via npm and run with npx ai-meta-mcp-server, or run with Docker using the provided Dockerfile. The server communicates via MCP stdio transport.

What is the purpose of this server?

It allows AI models to dynamically create and execute custom tools, extending their own capabilities through a meta-function architecture, while maintaining safety through sandboxing and human approval.