looper / Work / Fern
FFern fern.ph

Turning passive notes into interactive mastery.

Fern converts any topic or document into a structured, AI-generated study session, flashcards, quizzes, summaries, and graded problem-solving, then schedules it for long-term retention. We took it from concept to a live product.

Role
Design & buildCo-developed with LoopLabs Technologies
Category
EdtechConsumer AI web app
Platform
WebResponsive, desktop-first
Status
LiveShipped & in production
fern.ph/study/new
Fern dashboard. Start studying screen with topic input, time picker and task types
Live dashboard, fern.ph
The problem

Students drown in material and starve for retention. The default study loop, re-reading notes and highlighting, feels productive but barely moves long-term memory. The tools that do work (active recall, spaced repetition, varied task types) are tedious to set up by hand, so almost nobody does.

The brief we set ourselves: collapse the gap between "I have material" and "I'm actively studying it" to near zero, and make the resulting session genuinely adaptive, not a static deck of cards.

What we built

From input to a tuned session in seconds.

Six interlocking systems, each pulling its weight against the same goal: a session a student will actually finish and remember.

Dual input pipeline

Type a topic prompt or upload a PDF / DOCX. Content is auto-extracted and understood before a single task is generated.

AI session generator

The model analyzes the material, picks the best task types, and scales the whole plan to the time the student actually has.

Six task types

Flashcards, knowledge checks, quizzes, summaries, mnemonics, and graded problem-solving, routed intelligently per topic.

Focus mode

A distraction-free study surface that walks the student through tasks one at a time with live progress tracking.

AI answer evaluation

Open-ended problem attempts are graded by the model with specific, actionable feedback, not just right/wrong.

Spaced repetition

Due-card scheduling resurfaces material at the right interval so what was learned actually stays learned.

The stack

We chose a stack that's modern where it matters and boring everywhere else, fast to ship, cheap to run, and easy for one developer to maintain.

Next.js React TypeScript LLM API Document parsing Spaced-repetition engine Vercel
The outcome

A live product, not a prototype.

Fern shipped and runs in production today. The system holds together because every task type feeds the same retention loop.

6
AI-generated task types per session
2
Input modes, type a topic or upload a file
< 1 min
From input to a ready-to-run study session
PDF · DOCX
Document formats ingested and auto-extracted

Got a web app in your head? We've shipped ours, let's ship yours.

Twenty minutes to scope the build, talk stack, and get a fixed quote in your inbox by end of day.

Email us at [email protected]