Documentation

Mneva is a study app built around two ideas: get cards made fast, then review them at the right moment. You can write cards yourself, import them from Quizlet or Anki, or paste a PDF, lecture slides, YouTube link, web page, photo, or voice memo and let the AI write the cards.

Once a deck exists, Mneva gives you several ways to study it: classic flashcards, spaced repetition, an adaptive Train mode, practice tests, free recall, and six games. Every session feeds the Memory Map, which shows what you know and what is fading.

Mneva is free during launch. New accounts get the Standard plan, which includes AI generation, the full Memory Map, practice tests, the exam scheduler, and unlimited decks and folders. See Plans and limits.

1. 1Go to Sign up and enter your name, email, and a password.
2. 2If email confirmation is enabled, check your inbox and follow the confirmation link before logging in.
3. 3You land on the dashboard. The tutorial tour points out the main areas the first time.

Forgot your password? Use Forgot password. Mneva emails you a reset link that opens the reset page.

1. 1Open Create / Generate from the sidebar (or the Create button in the mobile bottom bar).
2. 2Pick a source tab: Paste Notes, Manual, Import, PDF, YouTube, PPT, Scan Notes, Scan Cards, Web URL, or Voice.
3. 3Add your material. For AI tabs, choose a card style (Q&A, Word:Def, or Mixed), a density (Essential, Balanced, Thorough, Exhaustive, or a custom count), and optionally a custom prompt.
4. 4Click Generate. Mneva names the deck from its content, saves it, and returns you to the dashboard.
5. 5Open the deck from My Library and choose a study mode.

Review AI cards before a serious study session. The deck editor lets you fix, delete, or add cards in seconds.

On phones the sidebar becomes a drawer. The bottom bar gives one-tap access to Home, Library, Create, Ranks, and More.

Source text is capped per generation (100,000 characters on the Standard plan). Long documents are trimmed; split very long sources into chapters for better coverage.

Card style controls the shape of each card. Q&A writes questions and answers. Word:Def writes term and definition pairs. Mixed uses whichever fits each fact.

Density controls how many cards the AI aims for from a given source:

Targets are capped by your plan (200 cards per generation on Standard). For large decks the AI runs a second pass that reorders cards into a teaching order and repairs formatting.

Definition length (Concise, Moderate, Detailed) changes how much explanation goes on the back of each card.

Custom prompt lets you steer the AI: "focus on dates and names", "write in Spanish", "use clinical terminology", and so on.

Math is preserved. Use $...$ for inline and $$...$$ for display LaTeX; the AI is told to do the same.

Quizlet. In Quizlet, open the set, choose Export, and copy the text. In Mneva, pick the Import tab, paste, and set the separators (tab between term and definition, new line between cards by default). Mneva shows a preview before saving.

Anki. Export the deck from Anki as an .apkg package and upload it on the Import tab. The file is read entirely in your browser. A field mapper lets you decide which Anki fields become the front and back; a heuristic pre-fills sensible defaults. Images and audio referenced by cards are extracted and kept. Cloze notes become fill-in-the-blank style cards.

• Media limits: 10 MB per file, 100 MB total per package.
• Both legacy and current Anki package formats are supported.
• Sub- and superscripts (chemistry, exponents) are preserved.

Standard. A front and a back. Both sides support bold, italics, lists, LaTeX math, and an image.

Fill-in-the-blank. On the Manual tab, click Fill in the blank on a card, enter the full sentence, and choose the word to hide. The front shows the sentence with a blank; the back reveals the word.

Image occlusion. Upload a diagram, draw rectangles over the regions to hide, and label each one. When studying, the image appears with the regions covered; tap each one to reveal its label. You can add occlusion cards while creating a deck or later from the deck's Occlusion editor. Ideal for anatomy, maps, and labeled figures.

Open a deck and choose Edit to see every card in a list. Use the search box to find a card by its front or back text, click a card to expand an inline form, and save. The bottom of the list has a New card form. Deleting a card asks for confirmation.

To rename a deck, change its description, or move it to a folder, use the menu on the deck card in My Library.

My Library lists every deck you own. Sort by recent, name, progress, accuracy, or times studied. Filter by public or private. Switch between grid and list views.

• Folders group decks. Create one from the collections panel, give it a color, then move decks in from each deck's menu. Folders also power the exam scheduler.
• Archive hides a deck without deleting it. Archived decks stay searchable and can be restored.
• Duplicate makes an editable copy, useful for forking a public deck you want to change.
• Delete removes a deck and all its cards permanently.

The three cards at the top of the library point you to a deck you have not studied yet, the one you studied last, and the one with the lowest accuracy.

The Mneva Chrome extension reads the page you are on and generates cards into a new or existing deck. Sign in from the extension popup with your normal Mneva account. Generations count toward the same monthly AI allowance as the web app.

Open any deck to start a flashcard session. Cards are shuffled. Flip a card to see the answer, then mark it right or wrong. When you finish, the results screen shows accuracy, time, XP earned, and the cards you missed, with a Review missed cards button to run through them again.

Bookmark a card to find it later, highlight text on a card, or report a public deck from the card header. The results screen also has tabs for the term list, games, and Luna.

Plain flashcard grading records your result but does not schedule reviews. Turn on SR mode for scheduling; see the next article.

Toggle SR in the study header to enable spaced repetition for a deck. Due cards come first, then a limited number of new cards, then anything else. After flipping a card you rate it:

Each button shows a preview of when you would next see the card. New cards go through short learning steps (10 minutes, then 1 day) before graduating to day-scale intervals up to a year.

Every card has a mastery status: New, Learning, Familiar (interval of a week or more), or Mastered (interval of three weeks or more). The results screen shows the distribution for the deck.

Mneva schedules with FSRS, a modern algorithm that models memory stability and difficulty per card, and keeps classic SM-2 data alongside it. The Scheduler tab in Settings lets you adjust the new-card limit per session and whether sessions default to due cards only. Tap the Why is this due? hint on a card for a plain-language explanation of its schedule.

The Review Hub collects due cards across every SR-enabled deck. The header shows how many cards are overdue or due today and a Start review button. Below it: a stacked bar of your overall mastery, a seven-day forecast of upcoming reviews, and a per-deck list sorted by due, mastery, or name.

Study all due (from the command palette or the hub) runs one mixed session across every deck instead of one deck at a time.

If the hub is empty, no deck has SR enabled yet. Open a deck, switch on SR, and grade a few cards.

Train gives every card a mastery score from 0 to 100, shown as five stages: New (0–20), Learning (21–50), Familiar (51–80), Proving (81–99), and Mastered (100). Your starting score comes from earlier Train sessions or, for new cards, from your past accuracy and speed. Lower cards come up sooner.

Mastery needs two kinds of proof. Recognition questions (multiple choice and true/false) can lift a card to 80 at most; a correct written answer is required to reach 100. Wrong answers cost more the higher a card is, but a card you have seen never falls below Learning, and it comes back within the batch.

Pick a goal before you start. Cram for a test is recognition-heavy and marks a card test-ready after one correct pass, with an optional weak-cards-only filter. Study to remember runs the full track, requiring recognition and recall. Quick review pulls only cards below Familiar.

Sessions run in focus batches of about seven cards (twelve when cramming). Finish a batch, see a short summary, and continue to the next. You can also cap a session at a number of cards or a time limit; progress is saved either way, so the next session picks up where you left off. The header shows the batch number and your overall mastery percentage.

Cards you already know well offer an I already know this shortcut on their first appearance: type the answer once to mark the card Mastered. Miss it and the card stays on the normal track. After a correct recognition answer you can tap Knew it instantly or Wasn't sure to fine-tune the jump.

Question options still apply: the question mix (mixed, multiple choice only, or written only), number of choices, strict spelling, hints, a per-question timer, direction, AI distractors, and Smart Grading. Answer wrong and an Explain this button asks the AI for a short explanation. At the end, Drill just these starts a focused batch over the cards you missed.

If you used Train before: the old Leitner boxes map onto the new stages (box 1 = New, 2 = Learning, 3 = Familiar, 4 = Proving, 5 = Graduated = Mastered), and your box progress was carried over as a mastery score.

Open a deck and choose Test. Pick which question types to include (multiple choice, written, true/false), the number of choices, the direction for each type, how many questions (10, 20, or all), and whether to shuffle.

All questions appear on one page with a progress bar. Submit when every question is answered. The results show a percentage and letter grade (A at 90% and above, B at 80%, C at 70%, D at 60%), a per-question review, and a Retry wrong button that rebuilds the test from the cards you missed.

Written answers use smart grading: a semantically equivalent answer counts even if the wording differs. Standalone tests can also be created from the Generate page (the AI Text, From Deck, Manual, and Import tabs under Tests) and saved to reuse.

Recall is a blank page. Type out everything you remember about the deck's topic, then submit. Mneva compares your writing to the cards and shows which concepts you covered and which you missed, so you can target the gaps.

Open the Games hub from a deck or from the results screen after a session. Games award XP and count toward streaks and achievements like any other session.

Luna appears on the results screen after a session and in Train mode. Ask her to summarize the deck, quiz you on the hardest ideas, or tell you which concepts you are likely to confuse. She answers only from the cards in the current deck, so she will decline unrelated questions.

When a session shows a pattern of missed cards, Luna can build a remediation deck: a small sub-deck aimed at exactly those concepts.

Luna is rate limited to 15 messages per minute and, like any AI, can make mistakes. Check anything important against your source.

Open a deck and choose Adapt. Mneva groups cards into struggling, at risk, strong, and unseen based on your sessions. Select struggling cards and run the diagnosis: the AI explains what makes each card hard to remember and proposes a rewritten front and back. Review each proposal side by side and accept or reject it. Accepted rewrites replace the card; rejected ones leave it unchanged.

Mneva is a progressive web app. On iOS, open it in Safari and choose Share → Add to Home Screen. On Android and desktop Chrome, use the install prompt or Install app in the browser menu.

Decks you have opened are cached for offline study. An indicator appears when you lose connection; progress syncs when you are back online.

The Memory Map draws your whole library as a solar system.

• Universe. One orb per deck, sized by card count and colored by memory health from strong (green) to critical (red); gray means never studied. The Today's priority panel lists up to 12 cards that most need review.
• Deck. Topic clusters orbit the deck. The AI groups cards by subject the first time you open a deck; use Re-cluster to regenerate. Each cluster shows studied count and mastery.
• Cluster. Individual cards in a spiral, each with a memory strength percentage. Click a card for six analytics: mastery, retention, accuracy, speed, attempts, and stability.

Memory strength blends accuracy (40%), how much has decayed since your last review (30%), repetition count (15%), and response speed (15%). The stats panel shows weighted mastery, coverage, and how many items need attention at each level.

Mneva extracts named concepts from your cards and infers relationships between them: prerequisite, contrast, part-of, builds-on, and related. The concept graph lays these out as a force-directed network.

Gaps appear as ghost nodes: concepts you probably need but have no cards for. They come from AI inference, from concepts referenced by other concepts but absent from your decks, and from curriculum templates. Each gap explains why it matters and suggests cards to fill it.

You can share a read-only snapshot of a map with a link from the map's share menu.

Finish any study session on a day and your streak grows by one. Miss a day and it resets to zero, unless you have a streak freeze. When you return after a missed day with a freeze available, Mneva asks whether to spend it and keeps the streak intact.

Freezes are granted monthly with your plan and earned at streak milestones. The Streak page shows a one-year calendar of your activity, your freezes, and the milestone ladder.

Streak days are counted in your local timezone. Optional email reminders warn you when a streak is at risk (Settings → Notifications).

Every correct card earns 10 XP. A perfect session adds a 5 XP bonus. Active streaks multiply the total: 1.1× at 3 days, 1.25× at 7 days, 1.5× at 14 days.

Your level is derived from total XP; each level takes progressively more. The dashboard shows XP to the next level, and levels unlock cosmetics in the rewards locker.

Achievements unlock for study habits, accuracy, volume, streaks, and specific modes. Each has a rarity: common, rare, epic, or legendary. The Achievements tab on your profile shows unlocked badges and a progress ring for each one still in reach.

The Rewards tab is your cosmetics locker: card themes, titles, and avatar borders unlocked by level, achievement, or streak milestone. Equip a title to show it next to your name; equip a card theme to restyle your study cards. The avatar editor lets you customize your avatar itself.

The leaderboard has four tabs: XP ranking, streak kings, card masters, and this week. A podium shows the top three; your row is highlighted. Names link to public profiles. Tier badges run from Novice to Legend.

Your rival is the person just ahead of you on the weekly board. The dashboard shows a head-to-head card so you can see what it takes to pass them.

Prefer privacy? Turn on Private profile or turn off Share study stats in Settings → Privacy.

Analytics shows session history and learning trends across all decks. Your Profile adds five tabs: Stats (totals and accuracy), Progress (weekly chart and a 17-week activity heatmap), Activity (recent sessions), Achievements, and Rewards.

1. 1Put the decks for the exam in one folder.
2. 2Open Exams, click New exam, give it a name and date, and choose the folder. One exam per folder.
3. 3Open the exam for today's plan: each deck in the folder with its due and unstudied cards, mastery percentage, and a link that opens it in SR due-only mode.

The plan moves through phases as the date approaches: first pass (get through everything), review (spaced reviews), and final (hit the weak spots). Urgency badges on the exam list run from On track to Today!. The dashboard shows your nearest exam.

Exam Hubs are collections Mneva maintains for specific exams. Each hub is organized by unit or topic following the exam's official framework, and every deck cites its source (OpenStax, Wikipedia, and other openly licensed material) with the license shown.

Fork any hub deck into your library to study it with SR, Train, tests, or games. Hubs are readable without an account; sign in to fork or favorite.

Make a deck public from its menu in My Library. Public decks appear in the Marketplace and in search, and get a shareable page anyone can open.

• Fork copies a public deck into your library so you can study and edit it. Mneva remembers what you have already forked.
• Favorite saves a deck to revisit later without copying it.
• Rate a deck out of five after studying it.
• Report a deck that is inaccurate or inappropriate from the study header.

Filter the Marketplace by category and difficulty, and sort by newest, popular, rating, or card count. Each deck shows a quality score built from evidence coverage, learner outcomes, freshness, creator verification, and reports.

Instant Study generates a 20-card set from a YouTube link, web URL, pasted notes, or a PDF up to 10 MB, with no account. The result lives at a short link you can share. It includes flashcards, matching, multiple choice, and typing modes.

Limits: five generations per day per network, at least 100 characters of source text, and up to 50,000 characters used. Identical sources reuse the same set for 90 days.

Signed-in users can click Save to my decks to claim the set as a normal deck with full SR and editing.

Your public profile shows your display name, username, level, equipped title, and (if enabled) study stats. Other learners reach it from the leaderboard or from decks you have published. Control visibility in Settings → Privacy with Private profile and Share study stats.

When you generate from a document, Mneva splits it into segments with stable IDs and records which segment each card came from. A deterministic check compares the card against its cited segment.

Open the provenance panel on a card to read the evidence excerpt, see the method (direct, inferred, or synthesized), and take review actions such as edit, regenerate, split, or merge. Model confidence is shown as a prioritization hint; it is never treated as proof.

The Trust Center explains the policy in full, including how reports are handled and how source licenses work.

The Sources page lists every document, page, or transcript you have generated from. Open one to see its segments and which cards cite each. If a web source changes, Mneva can detect it and mark dependent cards as stale so you can regenerate them.

Mneva is free during launch. Every new account is on the Standard plan at $0 per month:

Standard also includes the full Memory Map, all card types, web URL input, adaptive AI, practice tests, the exam scheduler, and every game. Check current usage in Settings → Usage. Allowances reset monthly.

Earlier subscribers on legacy Free, Scholar, or Scholar Pro plans keep those plans; manage them in Settings → Billing, which opens the Stripe customer portal. Cancelling keeps access until the end of the billing period. See Pricing for the latest.

Display name, username, and avatar are edited from your Profile page rather than Settings.

• Every study mode works with the keyboard; see Flashcard mode for shortcuts.
• Mneva honors your operating system's reduce-motion setting. Animations, confetti, and the Memory Map's motion all respect it.
• Settings → Accessibility offers a dyslexia-friendly font and a high-contrast mode.
• A speak button on cards reads the text aloud using your device's voices.
• Touch targets are at least 44×44 px, and the app respects safe-area insets on notched phones.

Your decks, cards, and study history are stored in a PostgreSQL database with row-level security, so only your account can read your records. Traffic is encrypted in transit. Instant Study never stores raw IP addresses, only hashes used for rate limiting.

Only decks you publish, and profile details you choose to share, are visible to others. Source documents you upload are private to you unless you publish a deck built from them, in which case only the evidence excerpts cited by cards are shown.

To delete your account, go to Settings → Account → Danger zone. Deletion removes all decks, cards, sessions, and profile data permanently. Read the full Privacy Policy and Terms of Service.

• "Limit reached" when generating. You have used this month's AI card allowance. Check Settings → Usage; the counter resets monthly.
• A YouTube video will not generate. The video has no captions or transcript. Try a different video, or paste your own notes.
• A PDF produces no cards. The PDF is a scan without a text layer. Use the Scan Notes tab with photos of the pages instead.
• Cards are wrong or off-topic. Review the source text that was extracted. Trim boilerplate, add a custom prompt, or lower the density.
• Streak reset unexpectedly. Streak days follow your device timezone. A session must finish (reach the results screen) to count.
• Cannot see a deck I forked. Forks are named "Title (Copy)" in My Library; search for the original title.
• The app feels slow or stale. Reload once to pick up the latest version; if installed as an app, close and reopen it.

Still stuck? Email support@mneva.app or use the feature request form for ideas.