Vocabulary Flashcards from Videos: Build Your Own Deck Without Manual Entry (2026)

You watch a movie scene, hear a word you don't know, look it up in your head, and keep watching. By tomorrow, the word is gone. By next week, you couldn't tell anyone you ever met it.

That gap — between seeing a word in context and actually remembering it — is what vocabulary flashcards are supposed to bridge. The problem is that the traditional way of building them, by stopping the video, opening Anki, typing the word, finding a sentence, pasting it in, adding a screenshot, is so disruptive that almost nobody keeps it up. This is a guide to a different approach: building a flashcard deck directly from the videos you're already watching, with spaced repetition doing the rest of the work.

The scheduling algorithm: FSRS

FSRS — Free Spaced Repetition Scheduler — is a modern algorithm for deciding when to show you a flashcard again. Spaced repetition itself isn't new. The insight, established by Hermann Ebbinghaus in the 1880s and refined ever since, is that recall is strongest when you review a piece of information just before you would have forgotten it. Show it too soon, you waste a review. Show it too late, you have to learn it again from scratch.

The classic algorithm — SuperMemo's SM-2, the historical default in Anki — applies a fixed multiplier to your next interval based on your self-rating ("again", "hard", "good", "easy"). FSRS does something different: it builds a model of your forgetting curve from your real review history and adapts intervals to that. The practical result is roughly 30% fewer reviews for the same retention rate, which matters a lot when you have a deck of 1,000+ words and only 20 minutes a day.

In Linglass, FSRS is enabled out of the box — nothing to configure. Anki also supports FSRS (added in version 23.10, October 2023), but you have to enable it manually in the deck options and run an optimizer over your review history. Not hard, but you have to know the option exists.

The problem with manual entry

The reason most people who try to build a flashcard habit fail isn't motivation — it's friction. Picture the flow with a traditional setup:

1. Hear an unfamiliar word in a Netflix scene.
2. Pause the video.
3. Switch to a dictionary tab. Type the word. Read the translation.
4. Open Anki. Create a new card. Type the front. Type the back.
5. Go back to Netflix. Take a screenshot. Crop it. Save it. Drag it into Anki.
6. Find the sentence the word appeared in. Copy it. Paste it. Tag the card.
7. Return to the video. Try to remember what was happening.

That's at least seven context switches for a single word. Multiply by 10 words per episode, and a 45-minute show takes 90 minutes to "process". Almost everyone gives up after a week.

The actual learning was supposed to happen while watching, but instead almost all the effort went into bookkeeping. This is why the tooling matters more than the algorithm.

Where Anki fits and where it breaks

Honestly: Anki is an excellent tool, and if you're willing to invest the time in setup and manual entry, it will work. Its strength is universal flexibility — any deck of anything. For our specific job (cards directly from subtitles with full context), the win goes to a tool purpose-built for video.

From video to flashcard in two clicks

A better flow looks like this: you're watching, you see a word you don't know, you click it — a translation popup opens — you press "Save". That's the whole interaction. The video keeps playing, no tabs to switch.

In those two taps, the extension does the bookkeeping for you:

• Reads the surrounding sentence from the subtitle track and saves it as the card's context.
• Captures the original-language audio of that sentence so you can re-hear it during reviews.
• Takes a screenshot of the frame so the visual context is there too — face, scene, action.
• Generates a contextual translation using AI that takes the whole sentence into account, so "run" in I went for a run doesn't get translated the same as in the company runs three offices.
• Schedules the card with FSRS to come back at the right interval based on your past reviews.

Try this on a YouTube video now →

The card is on your phone or laptop tomorrow morning, waiting in your review queue with all the context preserved. If you've ever wished you could "save this moment" while watching, that's what this is.

This is how the Linglass flow works on both YouTube and on Netflix with dual subtitles. If you're new to learning a language by watching content, the companion guide Learn English from YouTube covers how to pick the right videos in the first place.

Why spacing matters (the science, briefly)

Hermann Ebbinghaus's forgetting curve, the one psychology textbooks love to print, shows roughly 90% of new information slipping away within a week if it's never reviewed. The spacing effect — discovered in the same set of experiments — is the counter-trick: review at expanding intervals (a day, three days, a week, three weeks, two months) and retention stays near 90% indefinitely.

Watching one video and hearing a word once is exposure, not retention. Watching the same word across three videos in two weeks is closer to retention, but it's accidental and slow. Spaced repetition compresses that into a deliberate schedule: the algorithm makes sure each word comes back at exactly the moment your brain was about to forget it, which is when re-learning has the highest payoff per minute.

Practical setup (5 minutes)

1. Install the extension from the Chrome Web Store. It also works in Edge, Brave, Opera, and Yandex Browser.
2. Open the popup. Pick your learning language (the one in the video) and your native language (what you want translations in).
3. Open any YouTube video or Netflix show. Subtitles in your learning language appear automatically; if you've enabled dual subtitles, your native-language line shows underneath.
4. Click the first 5–10 unfamiliar words you encounter. Each click opens the translation popup; hit "Save" and the card lands in your deck.
5. Open learn.linglass.app/study when you're done watching. Run your first review session — typically 3–5 minutes.

Common mistakes

Saving too many words at once. If you save 50 words from a single episode, you'll get a 50-card review queue tomorrow on top of your existing reviews. Five to ten per video is the sustainable rate. Skip the words you're 80% sure of — they don't need a card.

Skipping review days. FSRS depends on actually doing reviews when they're due. Miss three days and the algorithm has to rebuild your interval estimates; miss a week and you'll come back to a long, demoralising queue. Five minutes a day beats an hour on Sunday.

Using SRS for words you already know well. If a word came up 15 times in a season of a show, you already know it. Putting it in flashcards just wastes review time. SRS shines for words you've seen once or twice and need to lock in — that's the gap it was built to close.

Treating it as the only learning activity. SRS is for vocabulary retention, not for grammar, not for speaking, not for listening practice. It's one tool in the stack.

What this isn't

Flashcards from videos won't teach you grammar. They won't give you speaking practice. They won't fix a weak ear for the language. They do one thing well: they take the words you've already encountered in real content and make sure you don't lose them.

The full stack for learning a language from videos looks like watching with dual subtitles (the input) + saving words you don't know (the capture) + reviewing them with FSRS (the retention) + output practice with a tutor or partner (the active recall). This article covers the middle two pieces. The other two are on you.

Frequently asked questions

How do I make flashcards from videos?

Install a browser extension that captures words as you watch. On YouTube or Netflix, click a word in the subtitles to open a translation popup, then press "Save" to add it to your deck — the card comes pre-filled with the sentence the word appeared in, a screenshot of the frame, and the native audio. There's no manual typing: the source video does the work of producing the context, you just decide which words are worth remembering.

What's the difference between FSRS and SM-2 (and does Anki have FSRS)?

FSRS (Free Spaced Repetition Scheduler) adapts review intervals to your actual recall patterns, while SM-2 applies a fixed multiplier based on your self-rating of each answer. FSRS gets you the same retention with roughly 30% fewer reviews. Anki has supported FSRS since version 23.10 (October 2023), but you have to enable it manually in deck options — SM-2 is still the default. In Linglass, FSRS is enabled out of the box; there's nothing to configure.

Can I save words from YouTube and Netflix automatically?

Yes, but "automatically" needs to be unpacked. The extension does not pre-pick words for you — that would flood your deck with words you already know. What it does is reduce saving to a click on a word and one tap on the "Save" button in the translation popup. The sentence, screenshot, audio clip and translation are filled in by the extension; the choice of which words to save stays with you.

Do I need to type translations myself?

No. When you click a word in the subtitles, the extension shows a contextual translation in your native language using AI that takes the surrounding sentence into account — so "run" in "I went for a run" gets translated differently from "the company runs three offices". You can edit the translation if it's off, but in 90% of cases the default is what ends up on the card.

How many words per day should I save with spaced repetition?

Five to ten new words a day is the sustainable rate for most learners. Below that you make slow progress; above 15–20 the daily review queue grows faster than you can clear it and the system starts to feel like a chore. Better to save fewer words from videos you genuinely enjoyed than many from content you watched on autopilot — context is what makes them stick.

Quick recap

• Manual flashcard entry breaks the watching flow and almost everyone quits within two weeks. Fix the friction first.
• A click on the word + one tap on "Save" should be enough to capture sentence, audio, screenshot and translation.
• FSRS gives you the same retention as SM-2 with about 30% fewer reviews — meaningful at 1,000+ cards. Anki supports it as opt-in; Linglass has it on by default.
• Five to ten new words per video. Skip the ones you already half-know.
• Five minutes of reviews a day beats an hour on Sunday.

The thing you're trying to build is a habit where the video does the work of producing context, and the algorithm does the work of producing the schedule. Your job is just to keep watching things you actually want to watch.

Try Linglass free →