Drop a lecture PDF, study notes, or paste any text. Get atomic spaced-repetition flashcards — one concept per card, optimized for Anki, Quizlet, or any SRS app. The PDF is parsed in your browser; only the extracted text is sent for card generation.
Drop a PDF, or click to browse
Parsed locally in your browser. Or paste text below.
0 chars (need 100+, max 100,000)
Cards(target — actual count depends on density)
Frequently asked questions
How does it generate flashcards from a PDF?
The PDF is parsed in your browser using pdf.js — text is extracted locally, never uploaded. The extracted text is then sent to Gemini with a strict prompt to produce 10–25 atomic flashcards (one concept per card), each with a question on the front and concise answer on the back. Optional topic tags are added.
Is the PDF uploaded to a server?
No. The PDF is parsed entirely in your browser. Only the extracted text — not the file itself — is sent to the Gemini API for card generation. Lecture slides, study notes, and other personal documents stay on your machine.
Can I import the cards into Anki?
Yes. The CSV download produces an Anki-compatible file with three fields per row: Front, Back, Tag. In Anki, choose File → Import, select the CSV, and map fields accordingly. Use the tag column to add cards to specific decks. JSON export is also provided for Quizlet, RemNote, or custom workflows.
What kind of material works best?
Definition-heavy material — vocab lists, biology/anatomy, chemistry, history dates, language learning, programming concepts. Less effective on free-form essay material where atomicity is hard. Best results come from clean, structured PDFs (lecture slides, textbook chapters); scanned image PDFs without OCR'd text won't work because there's no extractable text.
What's an 'atomic' flashcard and why does it matter?
An atomic card tests exactly one concept. 'What does the hippocampus do?' is atomic. 'Name three brain regions and their functions' is not. Atomic cards are the foundation of effective spaced repetition — you can confidently mark them right or wrong, and your review schedule adapts to each concept individually.
How big a PDF can it handle?
Text up to 100,000 characters per request — roughly 30–50 pages of text-heavy PDF. Larger documents should be split into chapters or topics; that produces tighter cards anyway. The 100k limit comes from Gemini's context window economics, not a hard ceiling.
Why do I sometimes get fewer cards than I asked for?
The 'Cards' setting is a target, not a quota. If your material is short or the model judges that fewer atomic concepts are present, it will return fewer cards rather than padding with weak ones. Better to ship 8 great cards than 20 forgettable ones.
Can I edit the generated cards?
Not in this tool — yet. For now, export to CSV and edit in a spreadsheet, then import. Inline editing is on the roadmap. The flip-on-click view lets you spot-check cards before exporting.