Turn handwritten notes into flashcards: 4 methods compared

First: what makes a flashcard worth making

A good card asks one question and has one answer. Every method below produces bad cards if you feed it bad questions, so get this right before deciding how to make three hundred of them.

• One fact per card. If the answer has three parts, that is three cards. A card with a paragraph on the back cannot be marked right or wrong, so you will always mark it right.
• A question on the front, not a topic. "Osmosis" is a label. "What is osmosis the movement of, and from where to where?" is a card.
• Answerable in under ten seconds. Anything longer belongs in a written answer against a mark scheme, not on a card.
• No cues you will not have in the exam. Cards that only make sense in the order you wrote them are testing the order, not the content.
• Reversible where it helps. Definition to term as well as term to definition — especially for vocabulary, formulae and quotations.
• Your own words. A card copied verbatim from a textbook tests recognition of that sentence. A card written in your own phrasing tests the idea.

A rough count: a page of A4 notes usually yields eight to fifteen cards. If it yields forty, they are too small; if it yields three, they are too big.

Method 1: writing them by hand on index cards

The slowest method and, for material you do not yet understand, the best one — because the act of writing a card is itself an act of retrieval.

To write "what are the three ways heat is transferred?" on one side and the answer on the other, you have to decide what the question is, what belongs in the answer and what does not. That decision is the learning. It is why students who make their own cards by hand often do better than students who study a better-made deck someone else built. The time is not wasted; it is a revision session with a deck at the end of it.

The costs are real, though. Twenty to thirty minutes a page, no automatic scheduling, cards that get left in a bag, and no way to shuffle intelligently — you end up doing the whole stack every time, which wastes most of the effort on cards you already know. Handwritten decks also cannot be split across the week without you doing the splitting.

Method 2: typing them into Quizlet or Anki

Typing your own cards into an app keeps most of the thinking benefit of writing them and adds scheduling, shuffling and a deck that lives in your pocket.

Anki is built around spaced repetition: it tracks how well you knew each card and decides when to show it again, so a card you find hard comes back sooner. It is free on desktop and Android, and a one-off paid app on iPhone. The interface is dated and the settings are unnecessarily deep, but it is the most serious tool on this list and it is what most medical students use for a reason.

Quizlet is friendlier and quicker to start, with a large library of decks other people have made. The free tier covers basic study modes; the more adaptive modes sit behind a subscription. Be careful with shared decks — a deck labelled with your exam board may have been made by someone on a different specification, and errors in a stranger's deck are invisible until they turn up in your answers.

Typing is faster than writing but the gap is smaller than it looks: fifteen to twenty-five minutes a page once you include making each card properly. And the moment you copy a page verbatim into cards to save time, you have lost the benefit that made this method good.

Method 3: OCR the page, then edit the text

The middle road: photograph or scan the page with a scanner app to get text out, then cut that text into cards yourself.

It works well on typed handouts, textbook pages and genuinely neat handwriting, and it saves the typing without giving up editorial control. It works badly on the notes most students actually have. Handwriting recognition still struggles with cramped writing, arrows and boxes, diagrams with labels, subscripts and superscripts, and anything in a margin. Chemistry equations, maths working and labelled biological diagrams come out worst — precisely the content flashcards are most useful for.

The failure mode to watch is silent error. OCR does not tell you it guessed; it hands you a confident sentence with the wrong number in it. If you use this method, read every line against the original page before it becomes a card, and expect to spend ten to fifteen minutes a page doing it. Skipping that check is how a misread "1.5" becomes something you have then rehearsed forty times.

Method 4: photograph the page and get a study pack

The newest option: photograph the notes and have an app write the summary, key concepts, flashcards and quiz for you, then check what it produced.

This is what FacilStudy does — up to six pages in one scan, built for handwriting rather than for typed pages, with a review screen before anything is saved so you can fix the subject, the topic or a wrong card in a few taps. The free tier is five scans a month; Plus is £4.99 a month for forty, plus an exam planner, a schoolwork-only tutor and written answers marked against a rubric. The cards then feed a fifteen-minute daily plan that chooses across all your subjects. FacilStudy is pre-launch, so there is an early-access list rather than an app to download today.

The honest limits. You skip the thinking that writing a card gives you, so on a topic you barely understand this is the worst method of the four — read the notes first, then scan them. Any AI can misread a word or write a card that is technically true and useless, which is why the review screen exists and why skipping it defeats the point. And a generated pack is only as good as the notes: if the page is wrong, the cards will be wrong too, confidently.

Where it earns its place is volume. Forty pages across eight subjects two months before exams is a real situation, and hand-writing four hundred cards is not going to happen. Cards that exist beat better cards that do not.

One paragraph, four ways: a worked example

The method changes how the cards get made, not what a good card is. Here is one paragraph of realistic GCSE Chemistry notes and what each method should turn it into.

Rates of reaction — collision theory. Particles have to collide with enough energy (activation energy, Ea) to react. Higher temp → particles move faster → collide more often AND more of the collisions have enough energy, so rate goes up. Higher conc (pressure for gases) → more particles per cm³ → more frequent collisions. Catalyst = lowers Ea, not used up. Surface area: powder reacts faster than lumps. Rate = amount of product ÷ time, e.g. 20 cm³ of gas in 40 s = 0.5 cm³/s.

A paragraph of Year 10 chemistry notes, roughly as written

Method 1: by hand

Writing them forces the decisions. The arrow between "higher temp" and "rate goes up" hides two reasons, so the card asks for both; the calculation gets a number on the back, because a number cannot be fudged. The first four cards:

The rest of the paragraph becomes cards five to eight the same way; "collision theory" alone never becomes one, because it is a label, not a question. That editing is the half hour.

Method 2: typed into Anki or Quizlet

Typed, the same four cards gain two shapes: the reversible card ("activation energy" on one side, its definition on the other, shown in either direction) and the cloze card — a sentence with gaps, which suits any line that is really a rule: "Higher concentration means more particles per ___, so collisions are more ___."

The temptation is the paragraph card: "Collision theory" on the front, the passage pasted on the back. It cannot be marked, so it will be marked right. If you are pasting, stop and split.

Method 3: OCR, then edit

Through a phone scanner, this paragraph typically comes back with the arrows missing or turned into hyphens, "cm³" as "cm3" or just "cm", the subscript in Ea lost, and "40 s" as "40 5". The prose survives; the chemistry does not.

So the edit is not cosmetic: restore the superscripts and subscript, turn the arrows back into words, and check the calculation line against the page digit by digit. You end with the same four cards you would have written by hand; the method saves typing, not thinking.

Method 4: photograph it and check the pack

Photograph the page and a draft comes back in a minute or two: five or six cards close to the four above, plus — typically — one of each of the two things the review screen exists to catch.

• The card that is technically true and useless. "What is collision theory?" with "The theory that particles must collide to react" on the back. Nothing wrong with it, and nothing on it you could lose a mark for. Delete it, or rewrite it as the two-conditions card.
• The misread. If "40 s" was written untidily, the back may say 0.5 cm³ per minute, or show 20 ÷ 4. The number is the check, every time: a card you did not write, with a number you did not check, is the most dangerous object in a revision folder.

The four methods side by side

Times assume a normal page of A4 notes producing eight to fifteen cards, and include the checking, which is the part every comparison leaves out.

Which method suits which subject

Cards are practice testing, which Dunlosky and colleagues' 2013 review rated as high utility across ages and subjects. The question is not whether cards work for your subject but what a card can hold, and which way of making them fits.

Science and maths. Cards hold definitions, equations, units, standard values and the steps of a set method — the recall layer that must be automatic before a six-mark question is answerable. Hand-write cards for topics you are still learning; type the bulk into Anki once you understand them, because the volume is high and the scheduling matters. OCR is weakest here, because notation is exactly what it misreads, and a photographed pack needs every symbol checked. Cards cannot do the problem-solving: choosing the method for an unfamiliar question is practised with questions.

English Literature. Quotations are the obvious card, and the front should carry the job, not the quote: "a quotation showing Macbeth's ambition before the murder, with the technique" is a card; "Macbeth, Act 1" is a label. Writing quotations out by hand is the memorisation, so Method 1 suits them best. An annotated poem is the worst input any method has — annotations are arrows over text — so photograph class notes on themes and context instead. No card writes the paragraph; that needs timed answers marked against the mark scheme.

History. Dates, names, statistics, causes and consequences: the most card-friendly essay subject, and the one with the most cards, so typing into an app earns its keep. It is also where OCR and photographed packs work best, because history notes are mostly prose with few symbols. Two shapes matter most: the "name three" card (three causes, three consequences, three reasons the League of Nations failed) and the sequence card ("what came immediately after this, and why"). The cards are the ammunition; the essay is still practised by writing essays.

How to choose (and the sensible mix)

Most students should not pick one method. Match the method to how well you know the material and how much of it there is.

1. Topic you find genuinely hard: write the cards by hand, slowly. The half hour is the revision.
2. Topic you know reasonably well but keep forgetting details of: type them into Anki or Quizlet so the app decides when to show them again.
3. A stack of neat handouts: OCR them, proofread against the page, then cut into cards.
4. Six weeks of scrappy notes across eight subjects: photograph them, check the packs on the review screen, and spend the time you saved actually answering questions.
5. Any method at all, one rule: if you never got a card wrong, it was not a useful card. Cards you always know should be retired, not rehearsed.

And whichever you choose, the cards are the easy half. What decides whether they work is reviewing them on a schedule with gaps in it rather than doing the whole stack once and feeling finished — which is the subject of the next guide.