SRS Spaced Repetition

Wrong words auto-enter the SM-2 spaced repetition queue, pushed precisely by algorithm — LangKey's core differentiator from immersive translators.

SRS Spaced Repetition

SRS (Spaced Repetition System) is LangKey's core differentiator. Wrong words don't just flash by — they get scheduled for review precisely via the SM-2 algorithm.

What is SM-2

SM-2 is a decades-old spaced repetition algorithm used by Anki and others:

  • Correct → lengthen next review interval (1d → 3d → 7d → 30d ...)
  • Wrong → shorten interval (re-learn in 1d)
  • 4-grade rating: didn't know / wrong / hesitated right / fluent — affects interval growth

Where Wrong Words Come From

All practice modes auto-capture unfamiliar / wrong words:

ModeTrigger
Input ModeMisspelled, missing, or extra word
Option ModePicked wrong option
Fill-in ModeAnswer doesn't match original
Listening ModeMisheard / wrong pick
ShadowingPronunciation not scored, wrong words enter vocab
Dictation ModeAny wrong word
Selection TranslateActively saved words

When a wrong word enters VocabularyDB, it automatically records:

  • Original sentence context
  • Source URL (video / article)
  • Source platform (YouTube / Bilibili / TED / any web page)
  • CEFR level estimate

Review Panel

Open LangKey popup → SRS Review:

  • Due today: words SM-2 says need review today
  • Overdue: longer-overdue words get higher priority (avoid permanent forgetting)
  • New words: max 20 new words join the queue daily
  • 4 rating buttons: Again / Hard / Good / Easy

Cross-Device Sync

Pro with cloud sync enabled, SRS progress syncs in real time across all Chrome / Edge / Firefox — study on mobile, review on desktop.

Anki vs LangKey SRS

FeatureLangKey SRSAnki
Source of wrong wordsAuto-captured, no manual entryManual card creation
ContextAuto-records original sentence / URL / platformMust add manually
Review entryBrowser popupStandalone app
CEFR estimate❌ (extra plugin needed)
AlgorithmSM-2SM-2 (default)

Feedback

Report SRS rhythm or scoring issues on GitHub Issues.