The Problem: Every day, regular people sign rental agreements, gym memberships, and employment contracts without reading them. Why? Because they are full of dense, confusing legal jargon. Hidden deep inside this "fine print" are automatic renewals, high cancellation fees, and unfair penalties. Hiring a lawyer costs thousands of dollars, and reading 20 pages of boring text takes too long. People deserve an easy way to protect themselves.
Inspiration & Idea:
A close friend of mine recently signed a gym contract without reading it. When he tried to cancel it a month later, he was shocked to find a hidden $150 termination fee in the text. This made me realize how common this problem is. I wanted to build a simple, free tool that gives regular people immediate peace of mind before signing anything. That is how JargonBuster AI was born.
How I Built It
As a solo developer, I wanted to focus on speed and ease of use. I built the entire app using Python. For the frontend, I used Streamlit because it creates clean web dashboards instantly without complex code. I used PyPDF2 to open and extract text from uploaded contract PDFs. Finally, I used the official Google GenAI SDK to connect the text to Gemini 2.5 Flash, which acts as the intelligent "brain" that audits the contract.
Challenges I Overcame
Since this was my first time using the official google-genai SDK, I faced some errors with the client setup and managing the environment variables (.env) securely. I spent time reading the documentation and fixed it by implementing secure system checks (os.getenv). Parsing messy PDF formatting into clean Markdown was also tricky, but I solved it by structuring the AI prompt with strict header instructions.
Accomplishments That I'm Proud Of:
I am extremely proud that I successfully built and fully deployed a working AI application completely on my own as a beginner. Taking a dense 20-page document full of legal traps and breaking it down into a visual, 4-point risk assessment in under 5 seconds feels like absolute magic.
What's Next for JargonBuster AI:
In the future, I want to add:
- Multi-language support: Translating hard English contracts directly into localized languages (like Urdu, Spanish, etc.).
- Interactive Chat: Letting users ask follow-up questions directly to the contract (e.g., "Can I leave this contract early without paying?").
Built With
- python
- streamlit
Log in or sign up for Devpost to join the conversation.