Lernara is built from a few ideas working together: Elo ratings size up you and each card, adaptive schedulers decide what to show next, a prediction model estimates your chance on every card, and progress tracking shows how far you've come. Each tile below is one piece of the picture.
Lernara treats learning as a rating system. Both learners and cards carry Elo ratings โ every answer updates yours and the card's, just like chess. Get a card right, it drops a few points; miss it, it climbs.
How ratings move
1420
+8
vs
1502
โ8
A correct answer on a harder card moves both ratings. Easier cards move them less; surprises move them more.
What you get from it
Adaptive
The scheduler picks cards near your current rating โ not too easy, not too hard. You always study at the edge of what you know.
Self-calibrating
As more learners answer a card, its rating sharpens toward the true population difficulty. The library gets smarter with use.
Yours to shape
Create your own sets, import from CSV, study community sets, track mastery over time, and climb the rankings.
A set is a collection of cards on a topic โ "French Nouns", "Country Flags", "Organic Chemistry Reactions". Sets can be private (just for you) or public (browsable on Explore).
A few example sets
French Nouns
420 cards
Country Flags
195 cards
My Med School Notes
87 cards
Creating a set
Go to Create and add cards one by one, or click Import Set to upload a CSV. The expected columns:
Mastery
Mastery is the prediction model's estimate of your chance to get a random card from the set right. It generalizes: a few right answers on hard cards convince the model that similar cards are within reach, so mastery can spike fast โ an expert starting a set can hit high mastery after an hour. It also drifts down between sessions if you stop showing up.
Country Flags
92%
French Nouns
68%
GRE Vocabulary
34%
Organic Chemistry
12%
Coverage
Coverage is a literal tally: the share of cards in the set you've answered at least once and got right on your most recent try. Unlike mastery, it doesn't generalize โ every card has to be earned individually. The expert who hits high mastery in an hour will still spend weeks bringing coverage to 100%. Miss a card later and it drops out until you bring it back.
Country Flags
100%
French Nouns
50%
GRE Vocabulary
90%
Organic Chemistry
15%
A card is one question and its answer. When you study, the scheduler picks cards one at a time, asks you, scores your response, and updates both your rating and the card's rating.
The lifecycle of a card
Create the card
You're asked
What's the capital of France?
Ratings update
โ Paris
1420 +4
1200 โ4
Card fields
What you see when the card is shown.required
What you need to produce.required
Extra context revealed after you answer โ sentences, mnemonics.
Hints or notes shown alongside the answer.
Starting Elo difficulty. The system refines it with use.
What the rating means
A card's rating is the Elo of a typical correct answerer. Miss a card rated 1500 and it climbs; get it right and it drops. Over time, card ratings converge to the true population difficulty.
Capital of France
900
EasyCapital of Brazil
1300
MediumCapital of Bhutan
1700
HardCapital of Tuvalu
2100
ExpertEvery card has a type that decides how you're asked. You can toggle types in your study settings โ e.g., turn off multiple choice to force active recall.
The five card types
Multiple Choice
Pick the correct answer from a list of options. Fast โ good for early exposure.
Capital of France?
Typing
Type the answer character by character. Strongest recall signal.
"dog" in Spanish?
True / False
Decide whether a statement is true. Useful for fact-checking drills.
The Nile is the longest river.
Manual
Self-grade after seeing the answer. For open-ended or creative cards.
Recite the first verse of the Odyssey.
Recall
Self-grade on a four-button scale (Again / Hard / Good / Easy). Anki-style โ for cards where effort matters as much as correctness.
Conjugate ser in the preterite.
The scheduler decides which card to show next. Lernara ships five โ you pick one in your study settings. They all see the same rating data; they just interpret it differently.
The five schedulers
Default
Picks cards near your current Elo, weighted toward those you haven't seen recently. The best default for steady progress.
You want efficient, balanced study.
Random
Picks any card uniformly at random. Useful as a baseline or when you want broad exposure without adaptive narrowing.
You're reviewing for breadth, or testing a set.
Nightmare
Every card is the one you'd most likely miss. More punishing than Target Accuracy's Brutal preset โ there's no balancing toward an average.
You've mastered most of a set and want to drill nothing but the remaining gaps.
Leitner
Sorts cards into 5 boxes with growing intervals (1 / 3 / 7 / 14 / 30 days). Right answers promote a card up; wrong ones drop it back to box 1.
You want explicit, predictable spaced repetition โ the Anki-style approach.
Target Accuracy
Aims sessions at an accuracy you set. Blends your last 10 actual outcomes with the 9 queued predictions ahead, then picks whichever card keeps the running mean closest to your target.
Your sessions swing wildly between easy and brutal โ you want a steady difficulty band.
Studying happens on the learn page. Each card is one round โ read the question, answer, see the result, move on. The top bar tracks how the session is going and how the system thinks of you.
A card in the wild
A multiple-choice card from a Chinese vocabulary set. The question is the pinyin and character; the options are short English meanings. Picking the correct option logs an answer and moves your rating.
The top bar
Four numbers, always visible while you study:
1
Answered
Cards completed in the current session. Resets when you start a new session.
1316
Your rating
Your global Elo. Goes up when you answer cards correctly, down when you miss them. Harder cards move it more.
55.9%
Mastery
Your estimated chance of getting a random card from this set right, given everything you've answered. Rises as you study; drifts down if you stop.
73.5%
Predicted accuracy
How likely you are to answer this specific card correctly, based on your rating and the card's. Lower numbers mean the card is at the edge of what you know โ that's where learning happens.
During-session controls
Settings
Toggle card types on/off, switch schedulers, adjust the difficulty of the choices โ without leaving your session.
Session stats
See what you've answered so far this session โ accuracy, time per card, mastery delta โ and review individual answers.
Lernara tracks your progress in four ways: a daily streak for showing up, monthly rankings for activity, per-set leaderboards for skill, and trophies for milestones.
Streaks
A streak day is any day you hit your daily goal โ 10 cards by default. Day boundaries are UTC midnight. Miss a day and the streak resets to zero โ no grace period.
7 days
current streak14h 23m
Monthly rankings
Global top-10 lists, recomputed monthly. Rankings come in two flavors โ users (who's most active) and sets (which content is most played). Each is sorted three ways: sessions, plays (cards answered), and accuracy.
Top users ยท by plays
liang
4,280cards
furkan
3,915cards
anya
3,104cards
nilo
2,870cards
Top sets ยท by sessions
Country Flags
1,820sessions
French Nouns
1,542sessions
GRE Vocabulary
1,311sessions
Capitals
988sessions
Per-set leaderboards
Every set has its own leaderboard, ranked by your rating on that set โ not your global rating. The top three each session hold a competitive trophy (Champion, Runner Up, Third Place); fall below rank 3 and you lose it.
French Nouns ยท leaderboard
anya
1,820rating
liang
1,742rating
furkan
1,689rating
kira
1,571rating
nilo
1,502rating
Trophies
Three trophy families. Two are permanent โ once earned, they're yours. The third is competitive: held by one user, lost when someone takes your spot.
Player
Awarded for total cards answered across all sets. Six tiers from Novice to Supreme Champion.
Novice Player
100 cards answeredExpert Player
100,000 cards answeredSupreme Champion
100,000,000 cards answeredCreator
Awarded for cards in your sets answered by other users. Seven tiers from Novice Creator to Myth.
Novice Creator
others answer your cards 100รMaster Creator
1,000,000ร answered by othersMyth
100,000,000ร answered by othersLeaderboard
Held by the top 3 of every set leaderboard. Awarded automatically and revoked when you drop below rank 3.
๐ฅ Champion
rank 1 on a set leaderboard๐ฅ Runner Up
rank 2๐ฅ Third Place
rank 3