ProBridge AI

Inspiration: Facing 40-page contracts and unreadable prescriptions that eat hours and invite errors.

What it does: Drag-and-drop any doc or image; ProBridge spits out a plain-language summary, instant translations, quizzes, and a Q&A chat.

How we built it: Flask backbone, PyMuPDF extractor, in-memory OCR fallback, OpenAI for simplify/translate, lightweight React front.

Challenges we ran into: Noisy phone photos, keeping latency under 3 s, and encrypting user data without bloating the stack.

Accomplishments we're proud of: Full MVP in 24 h, flawless mixed-file handling, smooth pharmacy and legal demos, zero crashes.

What we learned: Tuned pipelines trump brute force, clear UX copy earns trust, early security wins pay off.

What's next for ProBridge AI: PowerPoint export, dark mode, voice read-outs, mobile capture, SOC 2 certification.

Built With

helped with the designs and the overall format of the project and made sure it looked professional. Also helped in creating the slides and also the script.

Saatvik Guduru
Hello.
We all equally contributed. I did the frontend with the HTML and the UI.

Alay Shah
I did the back-end python coding and helped Alay with a little bit of the front-end. Saatvik did a lot of the design and helped with presenting.

Abhiram Pulijala

Leave feedback in the comments!

Log in or sign up for Devpost to join the conversation.