About Mneva

We built the tool we needed ourselves.

Mneva is an AI flashcard platform built by students who were tired of typing. Every feature answers a single question: does this improve retention? If the answer is no, it does not ship.

How it started

Most flashcard apps make you type every card by hand. That is fine for ten cards. It is soul-crushing for two hundred. We built Mneva because we were students who needed to study from lecture slides, textbook PDFs, and YouTube recordings, and no existing tool could turn those into flashcards without hours of manual work.

The first version was crude: paste text, get cards. It worked just well enough that we kept using it. Then we added PDF uploads. Then YouTube transcripts. Then voice memos. Each addition came from a real study session where we hit a wall and said, “this should be faster.”

Spaced repetition came next. We started with a basic Leitner box system and later adopted the SM-2 algorithm, the same one Anki uses. We added four-button rating because binary right-or-wrong is not how memory works. You can almost remember something, or remember it easily, or completely blank. The algorithm should know the difference.

Today Mneva is a small, independent product. We answer support emails ourselves. We read every feature request. We use Mneva to study for our own exams. That is not marketing copy. It is the only quality control that actually works.

What Mneva does

AI generation

Drop in a PDF, a YouTube link, a voice memo, or paste raw text. Mneva reads your material, structures it, and produces complete flashcard decks in under a minute.

Spaced repetition

Rate every card Again, Hard, Good, or Easy. The SM-2 algorithm schedules each review at the optimal interval. Hard cards return sooner; mastered ones stretch to months.

Seven study games

Matching, Multiple Choice, Typing, Speed Round, True or False Blitz, Survivor, and Meteor Strike. Each game reinforces recall through a different cognitive path.

Memory Map

A 3D knowledge graph where every card is a node colored by mastery, from teal to rose. See exactly which topics are solid and which need attention before an exam.

What we believe

01

Memory is the whole point.

Every feature answers one question: does this improve retention? If not, it does not ship. That single rule has killed countless ideas that seemed fun in meetings but did nothing for actual learning.

02

No ads. Full stop.

Ads destroy focus when you are holding information in working memory. We charge for Pro features instead. That is the deal: transparent pricing, no gimmicks, no attention arbitrage.

03

The feature request form is real.

We read every submission. Many of the best features came from students saying “this would help.” The roadmap evolves. If you have an idea, send it.

04

We are still building it.

Rough edges exist. Some bugs got fixed. Others have not surfaced yet. Every update ships because a real student needed it, not because a roadmap said so.

Under the hood

These tools solve real problems. Not trendy, just reliable.

Next.js + Vercel

Server-rendered pages deployed globally. No separate backend.

Supabase

PostgreSQL with row-level security enforced at the database, not the app.

OpenAI

Prompt-engineered per source type. PDF, YouTube, and voice each get tailored extraction.

SM-2 + Leitner

Two spaced-repetition algorithms. We did not reinvent either.

Stripe

Handles subscriptions. We never touch raw payment data.

Get in touch

We are a small team. We answer support emails personally. If something is broken, confusing, or could be better, tell us.

Try it.

No credit card. No catch. Upload what you are studying right now. If it does not help, walk away. Nothing to cancel.