🏆 AI Track Winner: AI Flashcard Maker

AI-powered Adobe Express add-on that automatically generates, formats, and lays out beautiful flashcards from any document—built to save educators hours of manual prep work.
ReactJavaScript

Project Type

Hackathon Winner → Production Collaboration with Adobe

My Role

Team Lead & Full-Stack Developer

Current Status

AI Track Winner at Adobe Express Hackathon (3,345 participants)
July 2025 - Present
GenAIReactPythonJavaScriptFlaskGoogle CloudBabel

Inspiration

For teachers, turning teaching materials into flashcards often means spending hours coming up with questions, formatting layouts, and preparing files for print. Many educators lack the design expertise (or time) to create polished resources.

That’s why we built the AI Flashcard Maker for Adobe Express—so any educator can instantly create beautiful, ready-to-print flashcards on a platform built to simplify graphic creation.

What It Does

With automated text extraction, AI concept generation, and auto-layout, the add-on instantly transforms documents and notes into engaging, print-ready flashcards.

  • Extracts text from PDFs, PPTX, PNG, and JPG.
  • Uses GPT-4o-mini to generate high-quality Q/A pairs.
  • Automatically builds layouts and inserts flashcards directly into the Express canvas via the Adobe Add-On SDK.

How We Built It

Using pdfjs and tesseract, the add-on extracts text from common teaching file formats such as PDF, PPTX, PNG, and JPG. The backend—powered by GPT-4o-mini—detects key concepts in the extracted text and generates high-quality questions and answers.

With the Adobe Express Add-On SDK, we calculate layout, spacing, typography, and grouping rules to insert flashcards directly onto the page. Hosting the backend on Google Cloud enables the add-on to be safely distributed via a private Adobe Express listing.

Challenges We Ran Into

  • Packaging and bundling the add-on for production deployment.
  • Navigating unclear or incomplete Adobe Add-On API documentation.
  • Deploying the backend to Google Cloud and integrating securely.

Impact & Accomplishments

  • Reduced educators’ manual flashcard preparation time by 80% through AI-driven Q/A generation and layout.
  • Built and deployed a fully functional Adobe Express add-on despite ambiguous documentation and rapidly evolving SDK features.
  • Designed and implemented backend infrastructure on Google Cloud, learned and deployed from scratch.
  • Engineered a CI/CD pipeline using Babel and custom npm scripts to automate packaging and deployment.
  • Led a team of 5 from ideation → prototype → production collaboration with the Adobe Express Add-On team.
  • Worked closely with UX designers to build polished, accessible interfaces through iterative, user-centered development.

Special Thanks

Thank you to the Networking at Tufts Lab for providing the LLMProxy, made possible by their research paper LLMProxy: Reducing Cost to Access Large Language Models.

Tech Stack

  • Frontend: React, JavaScript, CSS, Adobe Add-On SDK
  • Backend: Python, Flask, Google Cloud
  • AI: GPT-4o-mini, LLMProxy, text extraction (pdfjs + tesseract)
  • Tools: Babel, Git, custom CI/CD scripts