Thomas Raw is a multi-featured travel assistant designed to help users plan their trips with ease. It leverages the reasoning capabilities of Gemini via LangChain and integrates seamlessly into a Streamlit frontend.

🔍 Key features: At present the MVP (minimum viable product) consists of basic features which provide the rudimentary roadmap for the trip, advanced integrations are coming soon!

• Mr. Thomas Chats: A natural language interface to generate destination guides, custom itineraries, and trip suggestions.
• Budget Buddy: Generates a realistic travel budget based on user-defined parameters like duration, destination, and travel style.
• Where Is It?: Users can upload a travel photo, and the app attempts to recognize the location or landmark using Gemini Vision capabilities.
• Insight Centre: Export chat history, budgets, etc. for easy sharing and documentation.

🛠 Tech stack highlights: This tech stack is one of the best ones I've ever encountered in terms of prototyping fast (took me <5 hours to get this up and running on the cloud). Highly recommend it to anyone building small AI apps fast for hackathons with not much importance/emphasis on the UI. The UI is not bad, it's just hard to customise it your own way.

• LangChain for building structured prompt pipelines and memory management
• Google Generative AI (Gemini 2.0 Flash) for both text and image-based intelligence
• Streamlit for rapid UI development
• Session state used to persist and synchronize user data across app pages