Turn a PDF into Anki flashcards
Upload a chapter, a paper or a textbook section and get question/answer pairs as a two-column CSV you can import straight into Anki.
Making cards by hand is the bottleneck of spaced repetition — most people quit while still writing the deck. The reading is done; the cards are not.
This tool extracts the text, then writes atomic cards: one fact per card, short answers, no card that depends on another card to make sense. You review them before importing, so nothing wrong ends up in your deck.
What the output looks like
Real output: 2512 characters of chapter text in, 25 cards out, 16.0 seconds. First four below, exactly as generated.
| By 1788, what share of French crown revenue went to debt service? | Roughly half |
|---|---|
| Which wars left the French crown with its debt? | The Seven Years' War and French participation in the American Revolution |
| Which principal direct tax were the First and Second Estates exempt from? | The taille |
| Which finance ministers proposed taxing the nobility and were dismissed or forced out within two years? | Turgot, Necker, Calonne, and Brienne |
Who it is for
- Med and law students making cards the week before an exam
- Turning one textbook chapter into the smallest reviewable cards
- Converting a paper's methods and results into self-test questions
Not a good fit: Cards that only make sense with an image or a formula - only text is read.
Three things that improve the result
- One chapter per upload. A whole book gets shredded and the cards lose each other's context.
- Cards are deliberately atomic: one fact, one short answer. Merge them in Anki if you prefer denser cards.
- Skim before importing. The model occasionally states an author's claim as a fact.
How it works
- Upload a PDF (a chapter at a time works better than a whole book).
- Switch the output to flashcards and press the button.
- Download the CSV and import it into Anki — front in the first column, back in the second.
Questions
What is the CSV format?
Two columns, quoted, UTF-8 with a byte-order mark: front, then back. Anki maps them to your note type on import.
How many cards does it make?
Twelve to twenty-five depending on how much distinct material the document holds. Dense factual text gives more; narrative gives fewer.
Are the cards checked?
They are written from the document only, but a language model writes them. Read them before you import — a wrong card is worse than no card.