Multi-Agent-System-using-LangGraph-MCP-Supervisor-Guardrails-HITL
A demo multi-agent travel-planning assistant built with LangGraph and MCP, showing a supervisor agent, input guardrails and human-in-the-loop approval around a FastAPI web frontend.
This project is a demonstration multi-agent system that plans travel itineraries using LangGraph and the Model Context Protocol. A supervisor agent coordinates the workflow, input guardrails validate user requests before they reach the planner, and a human-in-the-loop approval step lets a person accept or request revisions to a generated plan before it is finalized.
The repository includes a FastAPI web frontend for sending planning prompts and approving drafts, an example MCP server that answers weather queries as a sample domain adapter, and client helper code for talking to that server. It is run locally with a Python virtual environment, installing dependencies from a requirements file and starting the FastAPI application with uvicorn.