An agentic travel planner that functions as a hyper-personalization engine, designed to automate and elevate the travel planning experience. This project showcases a tool-augmented agent built with LangChain that leverages a RAG pipeline to create deeply personalized itineraries, cutting user research and planning time by over 80%.
- Autonomous Itinerary Generation: The core agent autonomously performs multi-step reasoning to generate complete, day-by-day travel plans from a single natural language prompt.
- Adaptive Recommendations via RAG: Leverages a Retrieval-Augmented Generation (RAG) pipeline to provide context-aware suggestions for attractions, dining, and activities that adapt to stated preferences and implicit user needs.
- Real-Time Tool Integration: Seamlessly integrates with 5+ external, real-time APIs and tools (e.g., Web Search, Weather, Maps) to enrich plans with live data, demonstrating a robust tool-augmented architecture.
- Hyper-Personalization: Goes beyond basic filters to consider budget, interests, travel style, and accessibility needs to create truly unique travel experiences.
- Conversational AI Interface: An intuitive, chat-based UI powered by Streamlit for natural interaction.
- Budget-Aware Planning: Provides cost estimates and recommendations tailored to low, moderate, or high budget levels.
- Cultural & Safety Insights: Offers location-specific tips on local customs, etiquette, and safety considerations.
- Core Agentic Framework: LangChain
- AI Architecture: RAG (Retrieval-Augmented Generation)
- Machine Learning: Scikit-learn (for preference modeling)
- Frontend: Streamlit
- LLM Integration: Google Gemini Pro (with support for local models via Ollama)
- Core Language: Python
- LLM API: Google Gemini
- Weather API: OpenWeather
- Web Search Tool: Integrated for real-time information retrieval.
- Maps/Geocoding API: For location data and distances.
- Recommendation APIs: For sourcing real-time hotel and restaurant suggestions.
- Python 3.8 or higher
- An active internet connection for API access
- Clone the repository:
git clone <repository-url>
cd travel-agent- Create and activate a virtual environment:
# On Windows
python -m venv venv
venv\Scripts\activate
# On macOS/Linux
python -m venv venv
source venv/bin/activate- Install dependencies:
pip install -r requirements.txt-
Create a
.envfile in the project root. -
Add your primary API keys and configuration. Note: Other tools like web search are integrated directly.
OPENWEATHER_API_KEY=your_openweather_api_key_here
GEMINI_API_KEY=your_gemini_api_key_here
LLM_MODE=google # Options: google, local
- API Key Sources:
- OpenWeather Key: openweathermap.org/api
- Gemini Key: makersuite.google.com/app/apikey
- Start the Streamlit app:
streamlit run app.py- Open your browser and navigate to the local URL provided (usually http://localhost:8501).
travel-agent/
├── .streamlit/
│ └── config.toml # Streamlit configuration
├── app.py # Main application file (Streamlit UI & agent logic)
├── requirements.txt # Python dependencies
├── README.md # Project documentation
├── LICENSE # MIT License
└── .gitignore # Git ignore rules
Contributions are welcome! Please fork the repository, create a feature branch, and submit a Pull Request with a clear description of your changes.
This project is licensed under the MIT License - see the LICENSE file for details.