"Pydantic" matching MCP servers:

Policy-based governance for AI agent tool calls. YAML policies, approval gates, risk assessment, and audit logging across LangChain, OpenAI, Anthropic, and MCP.

A FastMCP server that enables AI engineers to run a production-ready MCP endpoint with in-memory token-bucket rate limiting and strict Pydantic v2 input validation. It connects to Claude Desktop, Cursor, and custom agent runners over stdio for safe tool execution.

Enables MCP clients to interact with a local AI agent through Ollama, providing tools for checking weather, current time, and available models, while tracing calls to Langfuse.