Fineprint

Inspiration

Contracts are full of complex legal language that most people don’t understand. Important details are often buried in fine print, leading to bad decisions, hidden costs, and legal risk. We wanted to build a tool that makes contracts simple, transparent, and accessible to everyone — not just lawyers.

What it does

Fineprint is an AI-powered contract risk analyzer. Users can paste or upload a contract, and the system breaks it down into key clauses, explains each one in plain English, flags risky or unfair terms, and suggests safer alternatives. It helps users quickly understand what they’re agreeing to and avoid hidden risks.

How we built it

We built Fineprint using Next.js for both the frontend and backend API routes, and integrated Google’s Gemini model for contract analysis. The backend processes input text (or files like PDFs), sends it to Gemini with a structured prompt, and returns clean JSON containing clause breakdowns. The frontend then displays the results in a simple, readable format.

Challenges we ran into

One of the biggest challenges was ensuring consistent and structured responses from the AI. Sometimes the model would return invalid or messy JSON, so we had to build parsing, validation, and fallback logic. Another challenge was handling different input types like PDFs and making sure the extracted text was usable for analysis.

Accomplishments that we're proud of

We built a fully functional end-to-end system that can take real contract text and generate meaningful, structured insights. The ability to break down clauses, explain them clearly, and assign risk levels makes the tool genuinely useful. We’re especially proud of making something that feels practical and real-world ready in a short time.

What we learned

We learned how to design effective prompts for LLMs, handle unreliable outputs, and build systems around AI rather than just calling it. We also gained experience integrating APIs, structuring backend logic, and building a clean user flow from input to insight.

What's next for Fineprint

We plan to improve accuracy with more specialized models and better prompting, add clause highlighting directly in documents, and support more contract types like employment and SaaS agreements. We also want to add features like contract comparison, saving analyses, and real-time suggestions while editing contracts.