Deedly

Inspiration

Every party in a real estate deal checks something different. Buyers check things like model, price, and location, lenders check credit and financing, and title companies verify ownership. However, no one evaluates the full property risk in one place, so problems often come up late in the process or when it is too late. Risk hides in how property records, ownership patterns, and neighborhood signals relate to each other. Title insurance focuses on past ownership issues and home warranties focus on the structure itself, but we noticed that neither of them look at the broader network of signals around the property. We chose to build Deedly to combine these signals into a clear view and help people spot potential problems early on.

What it does

Deedly is a property trust and intelligence platform. A user enters an address and the system analyzes the property to generate an overall score, title health rating, risk breakdown, and personalized insights. Instead of showing raw datasets, the platform interprets property and neighborhood information and explains what it means. Users can ask questions through an AI copilot chatbot such as whether the property is a safe investment, and the system responds using the actual analysis data. Deedly also produces a printable Buyer Confidence Report that summarizes key risks and explanations. Users are able to easily click through all of the categories and have the freedom to go back whenever necessary. Additionally, Deedly goes beyond a single property record by analyzing surrounding patterns such as nearby properties, shared ownership, and violation clusters. The model also provides an anomaly score of if it differs from its network. This helps uncover risks that may not appear in one report but become clear when the broader property network is considered.

How we built it

We built the frontend using Nextjs and React. For the backend, we used both the FastAPI server framework and Python. Python handles the property analysis and the related scores. The backend processes address inputs, combine public datasets, and generate the trust score, risk categories, and summary results. We created multiple FastAPI endpoints to access publicly available datasets. The database stores analysis records and report data so users can revisit and export their Buyer Confidence Report. We made the UI/UX simpler to have a clean effect and is easier for the users to use.

Challenges we ran into

One of our biggest challenges was finding free and accessible datasets that were also meaningful. Many real estate and title related records are private, inconsistent across different locations or platforms, or difficult to access without paid services. It was also challenging to verify the AI responses, as the responses are specific to each address.

Accomplishments that we're proud of

We are proud that we combined raw property records and public datasets into a working system that gives useful information instead of separate data points. The platform turns the data into a clear trust score and simple risk categories that are easy to understand. We also brought together hidden risk signals from different sources and showed them in one place. Additionally, the AI copilot explains these risks using the analysis results so users get helpful insights without confusing technical details.

What we learned

We learned the importance of clear communication between the frontend and backend and gained hands-on experience working with APIs and real-time data processing. Additionally, we learned how to blend Nextjs with Python and how to generate natural prompts. Finally, all of our team members didn’t have much experience with GitHub, so collaborating on it without any major issues was a good experience.

What's next for Deedly

In the future, we want to integrate real title records, deed history, and lien data and expand the system to help predict closing delays. We also plan to improve the AI copilot and allow lenders and title professionals to use the platform directly. Our goal is to help real estate decisions identify risk as soon as possible so buyers and professionals can act earlier and make more well-informed choices.

Built With

• fastapi
• googlegemini
• networkx
• next.js
• numpy
• python
• react
• typescript