This tool provides a production-ready, framework-free agentic workflow designed for efficient travel planning. Built from scratch with Python and the Model Context Protocol (MCP), it features a custom lightweight MCP Client/Server and supports multiple LLM providers, including OpenAI, Anthropic, and Google Gemini. Users can interact via an interactive CLI or a modern Flask-based web UI, benefiting from integrated tools for flight search, car rental, weather forecasts, and simulated payments. The project emphasizes production readiness with structured logging, Pydantic validation, robust error handling, state management, performance caching, comprehensive testing, and Docker support, also serving as an educational resource with an fully annotated codebase.
Key Features
01Custom MCP Client/Server for standardized tool communication with multi-LLM support.
02Framework-free agentic architecture built with standard Python.
03Production-ready features including structured logging, Pydantic validation, error handling with retries, and Docker support.
04Integrated tools for flight search, car rental, weather forecasts, and payment simulation.
050 GitHub stars
06Interactive CLI and a modern Flask-based Web UI with search history and real-time status updates.
