Inspiration
Navigating the real estate world is often confusing, especially for first-time buyers, renters, or young investors. From legal jargon to investment decisions, people often rely on word of mouth or expensive consultants. I wanted to simplify that process using AI — by building a conversational assistant that provides instant, smart real estate advice, tailored to Indian users.
What it does
EstateGPT is an AI-powered chatbot that acts as a virtual real estate advisor. It helps users make smarter decisions when it comes to buying, renting, or investing in property — especially in the Indian market.
How we built it
The core of the project — powering the chatbot with conversational, intelligent responses to real estate queries. I used prompt engineering to guide the AI to provide more localized, specific answers.
💻 Python: Used to connect the frontend with the GPT-4 API and handle input/output processing.
🌐 Streamlit: A fast and beginner-friendly framework to create a clean web-based interface where users can type questions and view answers instantly.
⚙️ Prompt Templates: I crafted prompt structures for different query types — such as property investment advice, rental suggestions, legal document checklists, and area comparisons — to make GPT responses more relevant and useful.
Challenges we ran into
Narrowing down the idea into something I could build solo in two days
Getting GPT to return specific and localized answers without access to real-time data
Balancing simplicity and usefulness — I had to leave out features like live property data or maps to focus on the core chatbot functionality
Accomplishments that we're proud of
Built a Fully Functional AI Chatbot: Successfully integrated GPT-4 to provide real-time, intelligent real estate advice. It’s a fully working MVP that answers complex queries about real estate, tailored for Indian users.
Simplified Real Estate Decisions: Created an easy-to-use interface that makes navigating real estate decisions more accessible. Users can get instant advice without the need for consultants or browsing multiple sources.
What we learned
How to integrate GPT-4 into a real-world use case
Prompt engineering to get relevant, non-generic responses
Streamlit deployment and UI design
How to scope down a big idea into a simple MVP for a 48-hour hackathon
Log in or sign up for Devpost to join the conversation.