How to Turn a PDF into Anki Flashcards on iPhone (3 Ways)

Why PDFs are where studying stalls

Almost everything worth memorizing arrives as a PDF: lecture slides, textbook chapters, guideline summaries, language worksheets. The spaced-repetition part is solved — Anki's algorithm will happily keep facts in your head for years. The unsolved part is the conversion step: turning 40 dense pages into a few hundred well-formed cards is hours of clerical work, and it's exactly where most people quit.

That conversion step is what you can now do on the phone itself. Here are the three routes that actually work, with honest trade-offs.

Way 1: Generate cards from the PDF, on the iPhone

The most direct route: read the PDF inside Guru and generate cards as you go.

1. Import the PDF — from Files, iCloud Drive, or the share sheet.
2. Read in the built-in reader. Select a page or a passage and hit the AI action — the app drafts flashcards from that content, not from the model's general knowledge of the topic.
3. Review the drafts before accepting. Delete the trivia, tighten wording, merge duplicates. Ten seconds per card here saves minutes of wasted reviews later.
4. Study. The deck schedules with FSRS, and everything exports to .apkg whenever you want it in desktop Anki.

The reason to generate from the document rather than asking a chatbot to "make flashcards about the renin–angiotensin system" is grounding: cards sourced from your actual material match your course's terminology, numbers, and emphasis — and you can check any suspicious card against the page it came from.

Best for: studying on the go, slide decks, chapter-by-chapter reading. Weak spot: scanned image-only PDFs need OCR quality to be decent; try a cleaner source file if generation looks thin.

Way 2: Convert on the web, study on the phone

If the PDF is a 300-page book, the phone screen is the wrong place to triage it. The GuruOwl web converter takes the upload, generates a structured deck (it handles whole books, image occlusion for diagrams, and a medical-grade accuracy check), and gives you an importable deck. From there:

• study it on the web, or
• export the .apkg and import it on your iPhone with scheduling intact.

Best for: large documents, diagram-heavy material, building a deck once for the whole semester. Weak spot: it's a two-step flow — convert, then import.

Way 3: The manual route (still underrated)

For a small set of make-or-break facts — drug doses, formulas, irregular verbs you keep missing — hand-written cards are still the gold standard, because formulating the question is itself encoding. On iPhone that means: add a card manually, write the front as a question you'd actually fail, keep the back to one fact.

Best for: the 20 facts you absolutely cannot get wrong. Weak spot: doesn't scale to a 40-page lecture.

What makes a card worth reviewing (whichever route you take)

AI or manual, the same three rules decide whether a deck works:

• One fact per card. "List the branches of the external carotid" is a quiz question, not a flashcard. Eight cloze cards beat one list card.
• Cloze-first for dense material. "The rate-limiting enzyme of glycolysis is [...]" keeps the sentence's context and hides only the retrievable fact.
• Source-grounded. Every card should be traceable to a page you read. If you can't tell where a generated card came from, delete it — a wrong card rehearsed for months is worse than no card.

Common questions

Does it work with scanned PDFs? Text-layer PDFs work best. Scans depend on OCR quality — if selection feels broken in the reader, generation will struggle too.

How many cards per page is normal? Dense lecture slides: 3–8 good cards per page. If you're accepting 20 per page, you're memorizing filler.

Can I edit cards after generating? Yes — and you should. Editing is where a generated deck becomes your deck.

Is the deck locked in? No. Everything exports to a standard .apkg, so the deck outlives any single app.

← All posts