// case study
QuizyLite
A study tool that turns PDF highlights into recall cards linked back to the exact page.
Role
Solo: design, frontend, backend, deployment
Timeline
Nov 2025 - present, in private beta
Stack
Next.js · TypeScript · Tailwind · Express · MongoDB · Firebase
The problem
Studying from PDFs is passive: you read, you forget. Making flashcards by hand fixes that, but it's slow, and once a card exists it's disconnected from the material. When you get one wrong, there's no way back to the paragraph it came from.
What I built
QuizyLite folds card-making into reading itself. You highlight a passage while you read, and it becomes a recall card: a flashcard, a fill-in-the-blank, or a custom question. There's no separate authoring step.
Every card stays linked to its source. Get one wrong and you can jump straight back to the exact page and paragraph it came from, so a failed card turns into a focused review session instead of a shrug.
Mistake review and progress tracking close the loop: the app resurfaces what you miss and shows how your recall improves over time, behind per-user accounts.
How it went
- Highlighting while reading replaces the manual flashcard step entirely
- Wrong answers jump back to the exact source page, turning mistakes into review
- Live at quizylite.app, in private beta
Tech notes
Next.js + TypeScript frontend with Tailwind CSS and TanStack Query; an Express + MongoDB API with Zod validation and rate limiting; Firebase Auth; pdf.js for reading. Built with Claude Code and the agent harness.
Have a similar project in mind?
Tell me what you're replacing or building, and I'll tell you what it would take.
Let's talk