Pindu extends Anki's spaced repetition system with a reading engine to help you build comprehension and vocabulary. Prepare text from your Anki collection, then get credit for reading it.
Pindu integrates with Anki as an alternative review mode. Read passages at your level, either extensively or intensively, and get credit for all words automatically.
Navigate to AnkiWeb's shared add-ons
Click "Download" and copy the add-on code
Open Anki, go to Tools → Add-ons → Get Add-ons, and paste the code
Restart Anki and find Pindu in your tools menu
As you read, mark each word by recall confidence. Pindu commits the marks to your Anki collection as real SRS reviews. Vocabulary maintenance becomes a byproduct of reading connected prose.
Pindu prepares passages from your vocabulary, calibrated to the words you know and are learning. Dial difficulty up for intensive practice, down for extensive practice, or bring your own text.
Pindu's reader bundles fast dictionary lookups, translations, text-to-speech, an LLM-based chat, and one-click adds for new words you discover. You can resolve confusion and capture vocabulary without breaking your reading flow.
Pindu was built to complement Anki, not replace it. Alternate within a session, do one first and the other later, or skip one entirely: the work compounds into the same daily progress.
We see the Pindu approach as an exciting fusion of traditional language acquisition methods. It is heavily inspired by both (a) spaced-repetition flashcard systems (e.g., Anki) for efficient vocabulary acquisition and maintenance and (b) graded readers (e.g., Sinolingua's excellent Chinese readers) for personalized reading. Pindu takes advantage of modern computing horsepower and AI models to take these systems to the next level.
Anki is excellent at what it does, but isolated flashcard review has real limits as your deck grows. Other technology-forward solutions in the space cluster around augmenting the reading experience for existing texts (e.g., LingQ, ReadLang, Kimchi Reader) or improving the UI and UX of vocabulary review (e.g., Hack Chinese). For the research behind why SRS and extensive reading are complementary, see our research post.
Pindu currently supports Chinese only. We're planning support for Japanese, Korean, Spanish, and other popular languages based on user demand.
Which language are you learning? Email us to vote for the next language, or join our email list to get notified when new languages launch.
Pindu is built for conscientious self-learners who prefer direct approaches over gamification. If you're comfortable with Anki's spaced repetition system and want to extend it to reading practice, Pindu is designed for you.
Pindu focuses on measurable progress through Anki's proven spaced repetition system rather than points, streaks, or leaderboards. Your motivation comes from real reading fluency gains, not artificial rewards. If you're wondering how to integrate it with your current Anki habit, Getting Started walks through your first session in detail.
Pindu uses advanced language models to create reading content that naturally incorporates your due vocabulary words. It considers factors like your current level, the number of words you're reviewing, and your chosen topics of interest. Passages are calibrated to high vocabulary coverage — about 95% by default — so most words you encounter are ones you already know well, leaving room for incidental vocabulary acquisition through context. For a technical walkthrough of the full session pipeline, see How Pindu Works; for the research that motivates this design, see our research post.
We think so. Language modeling is the core competency of modern-day Large Language Models. In our experience, LLMs excel at generating "descriptively correct" text, i.e., choosing normal words for a situation and using them according to their normal senses. While you may not encounter archaic and rare usages unless you specifically ask for them, we see that as a plus for the vast majority of language learners.
Pindu (品读, pǐn dú) means something like "to savor reading" in Chinese. The character 品 represents "tasting" or "savoring," which neatly captures the sense of satisfaction we take from language learning. We have gone through a few naming iterations ("BeyondVocab" and "Mosaic") but Pindu feels right.
