At the hackathon, our team built EZLaw, a tool that helps people instantly understand what laws might apply to their situation. The idea was simple: instead of having to dig through endless legal documents, you could just type in what happened — for example, “I got into a car accident” — and EZLaw would return the most relevant bills and their details.

We made this possible by combining AI with government data. First, we used the Gemini API to interpret the situation and pull out the top three most relevant bills, formatted as billType.billNumber.congress. We then wrote a Python parser that split each of those into individual fields (type, number, congress). Using those fields, we called the Congress API to fetch bill metadata, including the official bill ID, the title, which chamber approved it, and the congressional session it came from. Finally, we displayed everything in a clean, user-friendly way so people could see how their situation mapped to actual laws.

Getting there wasn’t smooth. We actually started on a completely different project with a 27-year-old ex-Google/Amazon engineer who we thought would bring invaluable experience to our team. Instead, after we wasted about ten hours working on his idea, he quit and left us hanging. That forced us to pivot hard. Our first attempt at pulling legal data was through LegiScan, but we quickly realized its API wasn’t practical for real-time use — almost everything was cached and wouldn’t update for a week, which killed the reliability we needed.

At that point, we only had about eight or nine hours left in the hackathon. We had to scrap LegiScan and rebuild our entire pipeline around the Congress API, which turned out to be much more flexible and reliable. It was a scramble, but we pulled it together in time.

In the end, we delivered a fully working prototype that let users turn messy, real-world situations into a matched legal context in just a few seconds. Building EZLaw under so much pressure showed us how to adapt fast, cut scope when needed, and focus on getting the core experience right.

Share this project:

Updates