ClipSmart

Designed for STEM students, ClipSmart uses AI to detect what kind of text you’re copying whether it’s math, physics, or a chemical equation and automatically converts it into the proper LaTeX script!

Comment

Inspiration

As STEM students, we often have to copy problem statements from textbooks or PDFs into LaTeX for homework submissions. Professors usually give problem numbers, and the statements contain mathematical expressions. Regular copy-paste breaks formatting or requires tedious cleanup, which is inconvenient. We thought: there must be a smarter way for the clipboard to understand what I’m copying and convert it automatically.

What it does

Our smart clipboard helps STEM students copy problem statements, math expressions, and as an additional feature, addresses more efficiently. It detects the type of content being copied and:

  • Converts math expressions into LaTeX automatically.
  • Reduces the inconvenience of manually transcribing content from textbooks or PDFs.
  • Additionally recognizes physical addresses and generates Google Maps directions.

How we built it

  • Frontend: Java desktop app for lightweight clipboard monitoring.
  • Backend: Python with NLP models to detect math, text, or addresses.
  • API Integration: FastAPI connects the Python backend to the frontend.
  • Storage: AWS S3 for saving most recent outputs.
  • Web Dashboard: Built with Spring to display and manage recent outputs.
  • Deployment: The web service is containerized and deployed using Docker for portability

Challenges we ran into

  • Parsing math reliably across textbooks and PDFs
  • Balancing automation with user control to avoid errors in pasted content.
  • Intercepting clipboard operations without interfering with normal copy-paste.
  • Integrating Java frontend, Python NLP backend, FastAPI, and AWS storage seamlessly.
  • Implementing difficult features under strict time control

Accomplishments that we're proud of

  • Built a functional, lightweight tool combining NLP, cloud storage, and desktop interface.
  • Automated LaTeX conversion that accurately handles a wide range of math expressions.
  • Added a practical feature for physical addresses to save everyday inconvenience.

What we learned

  • How to apply AI/NLP to real-world productivity problems.
  • How to integrate multiple technologies (Java, Python, FastAPI, AWS) into a smooth workflow.
  • How addressing minor inconveniences can have a big impact on user experience.

What's next for ClipSmart

  • Support for more content types
  • Smarter formatting suggestions based on context.
Share this project:

Updates