How to make flashcards with math formulas: three methods compared

There are three honest ways to make flashcards with math formulas: handwrite them on index cards, type them into Anki using its built-in MathJax support, or generate them from your lecture notes with an AI tool that typesets the notation for you. Each trades authoring time against features differently — there is no single right answer. This guide covers all three, including the actual syntax you will need, then the card-writing principles that matter regardless of which tool you choose.

The reason this question needs a real answer is that math notation breaks most flashcard tools. An equation like $\int_0^\infty e^{-x^2}\,dx = \frac{\sqrt{\pi}}{2}$ cannot be typed into a plain text field, and a flashcard that displays mangled notation trains you to recall mangled notation. The three methods below all solve the rendering problem; they differ in what the solution costs you.

Method 1: handwritten index cards

The oldest method still has genuine advantages. There is no learning curve and no rendering problem — your hand writes $\nabla \cdot \mathbf{F}$ as easily as it writes anything else. For derivation-heavy courses there is a hidden benefit: your exam will be handwritten, and fluency in physically writing integrals, matrices, and Greek letters is a real skill that typing never trains. Index cards are also flexible in ways software is not; a quick sketch of a supply-demand diagram costs nothing.

The weaknesses are structural. Paper cards have no scheduler: the best you can do is a Leitner box, which approximates spaced repetition coarsely and depends entirely on your discipline. They cannot be searched, edited cleanly, backed up, or shared. And they do not scale — a semester of real analysis or econometrics easily runs past two hundred cards, at which point the shoebox stops being a system and becomes a pile.

Best for: small decks, derivation practice, and students who remember best by writing.

Method 2: Anki with MathJax

Anki deserves its reputation. It is free and open-source on desktop and Android, its scheduling algorithm descends from SM-2 — the spaced-repetition method with decades of use behind it — and, importantly for math students, it renders mathematical notation natively. Since version 2.1, Anki ships with MathJax built in, so no additional software is required. You type standard TeX between \( and \) for inline math, or \[ and \] for displayed equations, directly in the card editor:

Front:  State the derivative of tan x.
Back:   \(\dfrac{d}{dx}\tan x = \sec^2 x\)

The back of that card renders as $\frac{d}{dx}\tan x = \sec^2 x$ — proper typesetting, identical to what you see in a textbook.

The honest cost is authoring time, not difficulty. Learning the core TeX commands — \frac, \int, \sqrt, subscripts with _, superscripts with ^, Greek letters like \epsilon — takes an evening or two; the syntax is a bounded skill, not a wall. But every formula on every card must be typed by hand, and a deck covering one course typically runs sixty to a hundred cards. Budget several hours of authoring before your first review session. (Anki also offers an older LaTeX mode that renders equations as images via a local LaTeX installation; almost nobody needs it now that MathJax is built in.)

A note on Quizlet, since it is many students' default: its math input is limited. There is a symbols palette, but no full LaTeX rendering, and students in notation-heavy courses commonly fall back to pasting screenshots of equations — which cannot be edited, searched, or read comfortably on a phone. For a vocabulary course this does not matter. For calculus it does.

Best for: students who want full control, plan to maintain decks over multiple semesters, and are willing to pay the up-front authoring cost.

Method 3: AI generation from your notes

The newest option removes the authoring step. With ExamTeX, you upload your actual course materials — lecture PDFs, slides, Word documents — and get back a flashcard deck with the mathematics already typeset in LaTeX and reviews scheduled with SM-2 spaced repetition. The formula cards arrive looking the way they do in your textbook, without you typing a single \frac.

The honest framing: your work shifts from authoring to editing, and the editing is not optional. A generated deck is a first draft. You should expect to delete cards that are too easy, tighten prompts that test recognition instead of recall, and add the conditions of applicability your professor emphasized that the source notes glossed over. You know which three lectures the exam will lean on; no generator does. The time savings are real — reviewing eighty cards is much faster than authoring eighty cards — but reviewing is still work, and skipping it produces a worse deck than a careful Anki session would.

To see what generated math cards look like before uploading anything, the calculus and statistics example decks are a fair preview.

Best for: students with notation-heavy notes already in digital form, and anyone short on time before an exam.

How the three methods compare

	Index cards	Anki + MathJax	ExamTeX
Math notation	Handwritten	Typed in TeX, per card	Generated from your notes, editable
Spaced repetition	Manual (Leitner box)	SM-2-derived scheduler	SM-2 scheduler built in
Authoring time	Moderate, all manual	Highest — every formula typed	Lowest — review and edit
Learning curve	None	TeX syntax, an evening or two	None for notation
Scales past 100 cards	Poorly	Well	Well

Card-writing principles for math, whatever tool you use

The tool matters less than the cards. These principles hold across all three methods.

One fact per card

Split "state the formula" and "when do you use it" into separate cards. A card carrying three facts gets graded wrong when you recall two of them, and the scheduler cannot tell which fact you are forgetting. Small cards keep the feedback honest.

Always include conditions of applicability

A formula without its hypotheses is a trap waiting for exam day. The card "State L'Hôpital's rule" is incomplete unless the answer includes that the limit must have the form $0/0$ or $\infty/\infty$ and that $\lim f'(x)/g'(x)$ must exist. Math exams routinely punish condition-blind application of memorized formulas — write the conditions onto the card, or onto a paired card of their own.

The front should force retrieval, not recognition

A front that shows the formula and asks whether it is correct only tests recognition, which is the weakest form of memory. Compare:

• Front: "State the integration by parts formula, and how you choose $u$."
• Back:

$\int u\,dv = uv - \int v\,du$

Choose $u$ by LIATE — logarithmic, inverse trig, algebraic, trig, exponential — taking $u$ as the factor earliest in the list, so that differentiation simplifies it.

Producing that answer from a blank prompt is what the exam requires; nodding at it on a card is not. The same logic underlies why practice testing outperforms rereading — covered in depth in our practice testing guide.

Review on a schedule, not in a binge

The cards only deliver long-term retention if the reviews are spaced. Anki and ExamTeX handle the scheduling for you; with paper cards, you are the scheduler. Either way, ten minutes daily beats two hours the night before — the evidence is summarized in our spaced repetition guide.

Finally, remember what flashcards are for. They handle the recall layer — formulas, definitions, conditions. The procedural layer still needs timed work on full problems, which is a separate tool: printable practice exams with answer keys pair naturally with whichever deck you build.

The bottom line

Handwritten cards cost nothing to learn but do not scale and leave scheduling to you. Anki renders math beautifully through built-in MathJax and schedules reviews well, at the price of typing every formula yourself. ExamTeX generates typeset cards directly from your lecture notes, shifting your effort from authoring to editing. Pick by deck size and available time — then apply the same principles regardless: one fact per card, conditions always included, and a front that makes you produce the math, not merely recognize it.