Complete Guide

All Levels • 9 min read

PDF to Anki Cards: Turn Any PDF into a Deck

Copying vocabulary out of a PDF by hand is the slowest part of studying with spaced repetition. This guide shows how to upload a textbook chapter, scan, or handout to NextLang and download a ready-made Anki deck, plus how to steer the AI so the cards match what you actually need to learn.

PDFAnkiAI GeneratorImportLanguage Learning

By Siarhei HamanovichPublished August 10, 2026

Table of Contents

What kinds of PDFs work

Reading a chapter twice feels productive, but the testing effect says otherwise: in Roediger and Karpicke's experiments, learners who practiced retrieving material remembered far more a week later than learners who re-read it. Turning the PDF you just read into flashcards converts passive review into retrieval practice, which is where the retention actually comes from. Source: Roediger and Karpicke (2006)

The generator does not just read the text layer. It also looks at a rendered image of every page, which is why the answer to almost every format question is yes:

Digital PDFs with a text layer

Coursebooks, vocabulary handouts, articles, and e-books exported as PDF. The generator reads the text directly, so these give the cleanest results.

Scanned pages

Scans of textbook pages or worksheets work even when the PDF has no text layer at all. The AI reads a rendered image of every page, so no separate OCR step is needed.

Photos saved as PDF

Phone scans made with Notes, Google Drive, or a scanner app arrive as photo-PDFs. They are read the same way as scans, page by page.

Slides and screenshots

Lecture slides or screenshots exported to PDF work too. Mixed layouts are fine: the generator picks out the vocabulary rather than copying the layout.

Two limits to know before you upload: a PDF can be at most 10 MB and 30 pages. Anything over that is rejected during validation, before a credit is spent.

PDF to Anki deck, step by step

Open the Anki generator

Go to the NextLang Anki generator and sign in. PDF upload is a premium feature, so make sure your account has credits.

What to do:

Free accounts can use Describe mode without credits; uploading files needs credits
One credit is spent per generation, whatever the card count
The same PDF flow works in the Quizlet, Mochi, and Brainscape generators

Switch to Upload file and drop in your PDF

In the Content section, pick Upload file instead of Describe, then drag the PDF into the dropzone or click to browse.

What to do:

Limits: up to 10 MB and up to 30 pages per PDF
Password-protected PDFs are rejected before anything is charged
You will see the file name and size in a chip once it is attached

Set languages, level, and card count

The dropdowns decide the languages of your cards, whatever language the document itself is in.

What to do:

Source language is what appears on the Front, target language on the Back
Pick a CEFR level from A1 to C1 so the vocabulary matches your stage
Choose 5 to 30 cards; the free cap is 15, premium raises it to 30
Use the prompt as extra context, for example: focus on the cooking vocabulary in chapter 3

Generate and review the cards

Click generate and the AI reads the whole document, then writes vocabulary cards from what it found.

What to do:

Expect cards built around useful vocabulary, not verbatim copies of sentences
Check the preview: you can edit the deck name and the cards before exporting
If the topic was too narrow, regenerate with a more specific prompt

Download the .apkg and open it in Anki

Export the deck as a native Anki package. Double-click the file and Anki imports it with fields and tags intact.

What to do:

The .apkg is built in your browser and downloads instantly
Nouns keep their articles and accents; tags become native Anki tags
Prefer CSV? The same result can be copied as a separator-delimited list

Troubleshooting common issues

If a PDF does not turn into the deck you expected, one of these fixes usually solves it.

The file is rejected before generating

Solution: Check the limits: 10 MB and 30 pages. Split a long book into chapter PDFs or compress the scan and try again. A rejected file never costs a credit.

The PDF is password-protected

Solution: Remove the password first: open the PDF, print it to a new PDF file, and upload that copy. Protected files are rejected during validation, with no credit spent.

The cards feel generic instead of matching the document

Solution: Add a focusing prompt such as: only the legal vocabulary from the contract, or: the food terms from pages 2 to 4. The prompt steers which part of the document becomes cards.

The cards came out in the wrong language

Solution: The language dropdowns always win over the document language. Set the source language to the language you are learning and the target to the language you want translations in.

The scan is blurry and results look thin

Solution: Re-scan at a higher resolution or take the photo in better light. The AI reads a rendered image of each page, so what a human cannot read, it cannot read either.

Getting focused cards from long documents

The difference between a mediocre deck and a great one is rarely the PDF. It is how you scope the generation.

Scope the document with the prompt

A 30-page chapter contains far more than 30 cards worth of vocabulary. Tell the generator what matters: the dialogue vocabulary, the verbs in the grammar boxes, or the terms a B1 learner would not know yet. One sentence of scoping beats regenerating three times.

Split long material by theme, not by page count

Two focused generations from a restaurant chapter and a travel chapter beat one generic pass over both. Thematic decks are also easier to study, because context clusters help recall.

Match the level to the material

Set the CEFR level to your level, not the document's. From the same B2 article, an A2 setting extracts the everyday words you still miss, while a C1 setting digs out idioms and low-frequency vocabulary.

Run the same PDF in both directions

Generate Spanish to English for recognition, then English to Spanish for production. Two small decks from one document train both skills without doubling your reading.

Workflows: textbooks, scans, and exam prep

The textbook chapter routine

After finishing a chapter, export or scan it, generate 15 to 20 cards at your level, and review them the same week while the context is fresh. The chapter becomes a deck in under two minutes.

Real-world documents

Menus, brochures, rental contracts, and letters are the vocabulary you actually need abroad. Scan them to PDF with your phone and turn each into a small themed deck.

Exam prep from past papers

Feed a past reading paper into the generator at your exam level. The unfamiliar words it extracts are exactly the gap between you and the pass mark.

One PDF, several apps

The same upload works in the Mochi, Quizlet, and Brainscape generators, and the all-in-one page carries your PDF and settings across platforms. Build the Anki deck first, then switch tabs and export for a friend who studies elsewhere.

Premium Content

Unlock the complete guide with all advanced techniques

Premium access includes:

Complete guides with all sections unlocked

Import-ready files: .apkg and .mochi packages, plus CSV, TSV, TXT and MD

Bigger sets: up to 30 cards per generation

Upgrade to Premium Browse Free Guides

FAQ

PDF to Anki: common questions

Quick answers about limits, scans, credits, and export formats.

Is turning a PDF into Anki cards free on NextLang?

File upload, including PDF, is a premium feature: you need credits, and one credit is spent per generation. Describe mode, where you type or paste your topic, works on free accounts. If a PDF is rejected during validation, no credit is spent.

How large can the PDF be?

Up to 10 MB and up to 30 pages per file. Longer material works best split into chapter-sized PDFs, which also produces more focused decks.

Do scanned PDFs and photo scans work?

Yes. The AI receives both the text layer and a rendered image of every page, so scans, photo-PDFs, and screenshots work without any separate OCR tool.

Does it copy sentences from the PDF onto cards?

No. The generator treats the document as source material and writes proper vocabulary cards from it: one term per card, translated into your target language at your chosen CEFR level.

Which platforms can I export to from a PDF?

All four generators accept PDFs: Anki exports a native .apkg or CSV, Quizlet an import-ready list, Mochi a Markdown or .mochi package, and Brainscape a CSV. The all-in-one page carries the same PDF across platforms.

What about Word documents or EPUBs?

Upload accepts PDFs and images. For other formats, use Print to PDF or Save as PDF from the app that opens them, then upload the result.

Your next chapter is already a deck

Upload a PDF and download a ready-to-study .apkg in minutes.

Generate an Anki Deck Anki Deck Best Practices

CSV to Anki: a free converter in your browser

Already have the vocabulary in a spreadsheet instead of a PDF? Convert the CSV straight to an .apkg deck right here. The file never leaves your browser.

Open the CSV to Anki converter