Saksham-sudan/Multi-Agent-Tourism-System

# AI Tourism Assistant A multi-agent AI tourism system built with LangChain that helps users plan trips, check weather conditions, and discover tourist attractions through natural language conversation. **Created by:** Saksham Sudan (USN-1CR22IS129) **Assignment:** Inkle AI Intern Assignment ## Live Demo View Live App on Streamlit Cloud ## Features ### Multi-Agent Architecture - **Orchestrator Agent**: Main conversational interface that delegates tasks to specialized agents - **Weather Agent**: Handles weather-related queries using Open-Meteo API - **Places Agent**: Discovers tourist attractions using Overpass API - **Conversation Memory**: Maintains context across multiple conversation turns ### Technical Capabilities - Natural language understanding powered by GPT-4o (via GitHub Models) - Real-time weather data fetching - Tourist attraction recommendations for any city - Context-aware responses (e.g., "What's the weather there?" after mentioning a city) - Clean web interface built with Streamlit ## APIs Used All APIs used in this project are free and require no authentication: - **Open-Meteo API**: Weather forecasting and current conditions - **Overpass API**: OpenStreetMap data for tourist attractions - **Nominatim API**: Geocoding service to convert city names to coordinates ## Project Structure ``` ai-tourism-system/ ├── agents/ │ ├── __init__.py │ ├── langchain_agent.py # Orchestrator agent with memory │ └── sub_agents.py # Weather and Places specialized agents ├── tools/ │ ├── __init__.py │ └── custom_tools.py # API integration tools ├── .streamlit/ │ └── secrets.toml # Streamlit secrets (gitignored) ├── app.py # Streamlit web interface ├── main.py # CLI interface ├── requirements.txt # Python dependencies ├── .env # Local environment variables (gitignored) └── .gitignore …