Blog

Flashcards From Highlights: Turn Reading Into Recall

Updated · 17 min read

Flashcards From Highlights: Turn Reading Into Recall

Listen to this post

0:000:00

Share this post

Contents

The best flashcards from highlights are written from a short list of passages that matter. They are not made from every yellow line. Each card keeps a link back to the passage it came from. Pick the highlights whose answer you would otherwise have to look up again, rewrite each one as a question, and review the cards on a spaced schedule. If every card can reopen its original paragraph, you can check a card when it feels wrong. A highlight only shows you passed over a sentence. A card makes you retrieve it.

This guide walks through the full routine for students, researchers, professionals and anyone with a pile of saved articles, papers, books or podcast notes. The same steps work for flashcards from notes.

TL;DR

  • Highlighting is not learning. Turn only the highlights you would pay a price to forget into cards, and aim for a small number per article.
  • One card, one idea. Write the prompt as a question, and keep the answer short enough to say in a breath.
  • Keep the source attached to every card, so a doubtful card sends you back to the exact passage instead of leaving you guessing.
  • Space your reviews over days and weeks. Research on retrieval practice supports testing yourself over rereading, and feedback after the test helps.
  • If you use AI to draft cards, treat the draft as a first pass. Write your own answer to the "why" questions before you look at it.

Table of Contents

Why highlights alone do not stick

Highlighting feels productive because it is easy. You read, something looks important, you mark it. The trouble is that marking a sentence tells your brain very little. Weeks later you have a page full of color and no memory of why any of it mattered.

Flashcards fix that by forcing retrieval. In a set of experiments published in 2007, Butler, Karpicke and Roediger found that material students had been tested on was recalled better after a delay than material they had only reread. They also found that giving feedback after a test improved retention compared with testing without feedback. That result maps neatly onto flashcards: the front is the test, the back is the feedback.

So the job is not to copy your highlights onto cards. It is to convert the passages worth keeping into questions you have to answer from memory.

Step 1: Decide which highlights deserve a card

The most common mistake is carding everything. A 30-page paper can easily produce 60 highlights. Sixty cards from one reading is a review pile you will abandon by Thursday.

Filter first. Here are three tests you can apply in seconds.

The cost test. Ask whether not knowing this will cost you real time later. Learning-system writer Matt Handzel uses a 5-minute rule: if not knowing something will cost you five or more minutes over your lifetime, it earns a card. A definition you will meet in every chapter passes. A statistic quoted once in the introduction usually does not.

The thesis test. Try to state the piece's core argument in one sentence. If you can, that sentence is a card. Handzel makes exactly one thesis card per article so the argument's structure survives even when the details fade. It is a good habit for papers, where it is easy to remember a figure and forget what the authors claimed.

The reuse test. Will you use this in an essay, a meeting, a lab note or a conversation? If you can name the occasion, card it. If you cannot, leave it as a highlight.

Three tests for selecting flashcard highlights

A useful target is a handful of cards per article, not dozens. Some readings will give you only two or three, and that is fine.

What to leave as a highlight

Not every marked passage should become a card. Leave these as highlights or notes:

  • Long examples and anecdotes. Keep one line on what they illustrate.
  • Statistics you can look up in seconds when you need them.
  • Passages you marked because they sounded good, not because you need them.
  • Anything you did not understand yet. Understand it first, then card it.

That last point matters. A card built from a passage you only half grasped will teach you to recite words, not ideas. If a paragraph is dense, get it explained before you card it. Omphalis, for example, explains a hard paragraph without taking you out of the document, so you can settle what a passage means and then decide whether it is card-worthy.

Step 2: Keep the highlight tied to its source

A card with no source is a claim you can no longer check. Months later you will wonder whether the number was 12 percent or 21 percent, and whether the author said "causes" or "is associated with". Without the passage, the card quietly turns into a rumor.

So store three things with every highlight before it becomes a card:

  1. The exact passage, as written.
  2. A short note in your own words saying why it matters.
  3. A way back to where it came from: the page, the timestamp or the link.

This is the design idea behind Omphalis highlights and notes that reopen the passage. Your personal note sits next to the original text, and from the note you can get back to the paragraph it refers to. When a card seems off, you check the paragraph instead of trusting your memory of it.

The same principle applies if your highlights start somewhere else. If you read on an e-reader, you can import Kindle highlights and keep them useful rather than leaving them in a long unsearchable list. If your PDF is a scan and the text will not select, OCR makes scanned text selectable so you can highlight it in the first place.

Step 3: Rewrite each highlight as a question

This is the step where a highlight turns into a flashcard. A highlight is a statement. A card is a prompt that makes you produce the statement.

Take this highlighted sentence from a reading on memory: "Retrieval practice, the act of recalling information from memory, strengthens later retention more than restudying."

A weak card copies it:

  • Front: Retrieval practice
  • Back: The act of recalling information from memory, which strengthens later retention more than restudying.

It looks tidy, but it is vague. The front does not ask anything specific, and the back holds two ideas.

Two stronger cards:

  • Front: What is retrieval practice?

  • Back: Recalling information from memory instead of rereading it.

  • Front: Compared with restudying, what does retrieval practice do to later retention?

  • Back: It strengthens it.

Splitting it gives you cards you can mark right or wrong without argument. That is the standard to aim for.

Rules for a good prompt

  • One idea per card. If the answer has an "and" in it, you probably have two cards.
  • Ask something with a definite answer. "Tell me about X" invites a fuzzy reply and a lenient grade.
  • Include enough context. A card that reads "What was the sample size?" is useless in a deck of fifty papers. Write "In the Smith 2019 sleep study, how many participants were there?"
  • Keep the answer short. If you cannot say it in a breath, shorten it or split it.
  • Avoid cues that give the answer away. If the question repeats the answer's key words, you recognize instead of recall.
  • Use your own words where you can. Quoting exactly matters for definitions and for legal or technical phrasing. Otherwise a paraphrase shows you understood.

Step 4: Pick the card type that fits the passage

Different highlights call for different formats. Long-form reading tends to produce four useful kinds.

Highlight type Card format Example front
Definition Question and answer What does "survivorship bias" mean?
Key term inside a sentence Cloze (fill in the blank) Spaced practice spreads reviews over ___ instead of one session.
Argument or claim Why or how question Why do the authors argue the first model fails?
Process or sequence Ordered steps List the four stages of the method in order.
Thesis of the piece One-sentence summary What is this article's main argument?

A cloze card hides a word or phrase in the original sentence. It suits facts and terms, and it is fast to make from a highlight because the sentence is already written. Use it for details. Use question and answer cards for ideas, where you want to explain something in your own words.

For papers, a good minimal set is the thesis card, one card for the method, one for the main finding and one for the limitation the authors admit. That is four cards, and it gets you through most conversations about the paper. If papers are your main reading, Omphalis has a page on how to read and understand research papers with explanations and notes in one place.

For talks and podcasts, a highlight is a moment, not a paragraph. Write the claim in a sentence, add the timestamp in the note, and card the claim. If you watch lectures, you can read YouTube talks with explanations and notes so the moments you save keep their context.

Step 5: Use AI to draft cards, then do the thinking yourself

AI can speed up the dull part: turning a list of highlights into draft questions. It is useful, and it is also where careless cards come from. A model will happily produce twenty crisp-looking prompts from a passage it half understood, and some will be subtly wrong.

A workable split comes from Handzel's system, which separates two kinds of cards. For plain factual cards, such as definitions and names, an AI draft with a light review is enough. For conceptual "why" and "how" cards, he has the AI show only the question, writes his own answer first, and only then compares it with the AI's answer. He argues that this is where real learning happens, because producing an answer before seeing one is itself a form of retrieval.

A short checklist for any AI-drafted card:

  • Does the answer match the source passage, word for word where it matters?
  • Is it one idea?
  • Could you answer it from memory, or does it need the page open?
  • Would you understand the front in six months with no other context?

That first question is why source-linked notes matter. When you can jump from a card to the original paragraph, verifying a draft takes seconds. Omphalis also lets you ask questions about what you are reading, with citations back to the source, which helps you confirm what a passage says before you commit it to a card.

Step 6: Review on a spaced schedule

Making cards is half the work. The other half is when you review them.

The 2006 review by Roediger and Karpicke reports that spaced retrieval generally improves retention compared with massed practice, meaning cramming all your reviews into one block. The same review notes that evidence on whether intervals should expand over time or stay equal was mixed. So the safe takeaway is simple: spread your reviews out, and do not worry about finding the one perfect interval.

A practical routine:

  1. Review new cards the day after you make them.
  2. Review again a few days later, then about a week later, then every few weeks.
  3. Cards you get right move to longer gaps. Cards you miss come back soon.
  4. Spend ten minutes a day, not two hours on Sunday.

Most flashcard apps handle the scheduling for you. What they cannot do is decide whether a card was worth making. That is why the filtering step comes first. If you want a deeper look at scheduling, this overview of a retrieval strategy for learners covers how to space and structure recall. For a walkthrough that starts from your notes, see our spaced repetition from notes workflow.

Mark misses honestly

Grade yourself strictly. If you answered "sort of", it was a miss. And when a card keeps failing, do not just repeat it. Reopen the source passage, work out what you misunderstood, and rewrite the card. A card that keeps failing is usually a badly written card or an idea you never understood.

A worked example: one article, five cards

Suppose you read a long article on how sleep affects learning and highlight eighteen passages. Here is how the filter cuts that down.

  1. Thesis card. The article's argument in a sentence: sleep after studying helps consolidate what you learned, so cutting sleep to study more backfires. Card it.
  2. Definition card. One highlight defines memory consolidation. It will appear in every related reading. Card it.
  3. Mechanism card. A highlight explains why the process happens during sleep. You wrote a "why" note beside it, so you write the answer yourself first, then compare. Card it.
  4. Cloze card. A sentence names the stage of sleep linked to the effect. Hide the name. Card it.
  5. Practical card. A recommendation you might use during exam week. Card it.

The other thirteen highlights stay as highlights: two anecdotes, a handful of statistics you can look up and several sentences you marked because they sounded good. Five cards, each tied to its passage. A review pile like this takes a couple of minutes a day, and you will keep doing it.

Getting highlights out and into your card tool

Once you have chosen your highlights and written your prompts, you need to get them into your flashcard app. The simplest route is a table: one row per highlight, with columns for the passage, your note, the source and the question you wrote. Omphalis supports export, and the guide to exporting highlights to CSV shows a workflow that keeps the source with each row. Most flashcard apps can import a spreadsheet file, so the question becomes the front, the answer becomes the back, and the source column goes in an extra field.

For a student's view of the same routine, see the PDF highlighting strategy that turns highlights into flashcards.

Common mistakes

  • Carding while you read. You do not yet know what matters. Read first, mark lightly, then choose cards after.
  • Too many cards at once. If a session takes more than fifteen minutes, you made too many.
  • Cards that need the page. If you cannot answer without reopening the source, the card is too vague.
  • Never deleting cards. Retire cards for ideas you no longer need.
  • Skipping the "why" note. One line on why a passage mattered tells you, later, what the card should ask.

Where Omphalis fits

Omphalis is a platform for saving, reading and annotating articles, papers, podcasts and videos. It cleans and structures what you save, explains complex passages, and attaches your notes to the original text. That is the stage of the routine where most flashcards are won or lost: deciding what a passage means, what to keep and where it came from. You can import PDFs, EPUBs, Word documents and feeds, and export your highlights when you are ready to build cards.

The Free plan has no published price. Pro is $9 per month or $90 per year, Premium is $19 per month or $190 per year, and Scholar is $39 per month or $390 per year. Unit packs start from $10 one-off. Plans were read from the Omphalis pricing page in September 2026, which is also the place to check what each plan includes now.

Omphalis

If you want to save articles with their highlights and notes in one place, you can start on the Omphalis home page and, if you read in a browser, add the Chrome and Firefox extension.

Our perspective

Most reading tools are built around speed: shorter summaries, faster skims, more saved links. We think that gets the order wrong. A summary can leave you feeling you understand something you could not explain, and flashcards made from a summary inherit that gap.

That is why we build Omphalis around the full source. We explain difficult passages inline, attach your notes to the original text, and let you ask questions with citations, so the card you eventually write comes from something you understood. The question "what do I need to remember here?" is easier to answer when the passage is still open next to your own note.

Flashcards are a strong way to make reading last. The part that comes before them, reading carefully, marking selectively and noting why, decides whether the cards are worth reviewing. We want that part to be calm and fast, with the source always one click away.

Omphalis Team

FAQ

Can I make flashcards from notes instead of highlights?

Yes, and the method is the same. Pick the notes that capture an idea you will need again, rewrite each as a single question, and keep a link to the source. Notes in your own words often make better cards than copied highlights because you have already processed the idea.

How many flashcards should I make from one article?

Fewer than you think. A handful is a good target, and some articles deserve only two or three. Start with a thesis card, then add cards for definitions and the claims you will actually use. If a review session runs past fifteen minutes, you have too many.

Can ChatGPT or another AI make flashcards from my highlights?

AI can draft cards quickly, and that helps with factual ones. Check each draft against the source passage, since the wording can drift from what the author said. For "why" and "how" cards, try writing your own answer before reading the AI's, so the card trains your understanding and not only recognition.

Are flashcards an effective way to learn?

Research supports the testing behind them. In experiments from 2007, tested material was recalled better after a delay than reread material, and feedback improved retention. Flashcards work best when each card is clear and you review them over spaced intervals.

Sources