The honest answer to "what is the best notes app for learning a language" is that for the vocabulary half, a notes app is the wrong category. Words are a scheduling problem, not a storage problem, and the tool that solves scheduling is a spaced repetition system.
That still leaves three jobs a notes app does better than any flashcard app, which is why most serious learners end up running two tools rather than one.
Split the work first
| The job | The right tool |
|---|---|
| Remembering 3,000 words for years | Anki or another SRS — a scheduler, not a list |
| Catching a phrase you just heard | Any fast capture app — speed is the only requirement |
| Grammar you keep getting wrong | A notes app — one page per rule, in your words |
| Sentences you collected but haven't turned into cards | A notes app inbox — a staging area, cleared weekly |
| Speaking practice and mistakes | A notes app or journal — a running log |
The mistake almost everyone makes is keeping vocabulary in a note. A word list in a document gets reviewed twice, because nothing tells you when to look at it and the top of the list gets all the attention — the serial position effect in action. A scheduler shows you each word at the moment you are about to forget it, and the difference in outcomes is not subtle. That is the whole case for spaced repetition, and the distributed-practice literature behind it is one of the strongest results in learning research.
The shortlist
Anki is the default and still the best for vocabulary. It is free on desktop, Android and the web, and a one-time purchase of around $25 on iOS which funds the project. Modern versions ship FSRS, a scheduler that models your own forgetting rather than applying a fixed formula, which meaningfully cuts the number of reviews needed for the same retention. It is ugly, the card editor is fiddly, and the payoff is that it will still be there in ten years with your deck intact. Anki vs Quizlet covers the friendlier alternative and what it costs you.
Your existing notes app — Apple Notes, Obsidian, Notion, whatever — handles everything that is not a card: grammar explanations, a log of mistakes, sentences waiting to be processed, notes from lessons. Do not migrate for this. The requirement is one folder and fast search.
Obsidian earns a specific mention if you are studying a language with a writing system or a grammar you want to map: plain files, links between related rules, no subscription. See plain text notes.
A voice recorder is the underrated one. Recording yourself speaking for two minutes a week and listening back a month later is the clearest progress signal available, and it costs nothing. Voice note-taking covers the wider technique.
Sentence mining: the part that decides everything
The single biggest quality difference between language decks is not the app. It is whether your cards are words or sentences.
A card with perro = dog teaches you a mapping that falls apart the moment the word appears in a real clause. A card built from a sentence you actually encountered — where you understand everything except one element — teaches the word plus its grammar, its register and a context to retrieve it from. That extra context is a second route back to the memory, which is the mechanism in context-dependent memory, and it is why sentence cards beat word lists even though they take longer to make.
The practical loop:
- Capture the sentence where you met it, verbatim, with the source. Speed matters; use whatever is fastest on your phone.
- Leave it in an inbox, not in the deck. Half of what you catch will not be worth a card.
- Once a week, turn the survivors into cards. Making the card is itself a study session — the generation effect means the words you build cards for are already half-learned.
- Delete the rest without guilt. See when to delete notes.
Step two is the one your notes app is for, and skipping it is why so many decks are full of words nobody wanted.
The grammar notebook
Grammar is the opposite of vocabulary: a small number of rules, each of which needs to be understood rather than memorised, and each of which you will get wrong for months. This is genuine notes-app territory.
One note per rule. Your own explanation, not the textbook's — if you cannot write it in three lines, you have not understood it, which is the Feynman technique doing its job. Then, underneath, a growing list of your own real mistakes with the correction. That mistake log is the most valuable document in the whole setup, because it is the only one that is specifically about your errors rather than the language in general.
The two failure modes
Collecting instead of reviewing. A thousand saved words and no review schedule is not learning, it is an archive with a language theme. This is the collector's fallacy, and it is unusually easy to fall into here because saving a word feels like learning it.
Reviewing instead of using. The opposite failure, and more common among people who like systems. Cards are a supplement to reading, listening and talking to people, not a substitute. If your deck is at 4,000 cards and you have not had a conversation, the tooling has become the hobby.
What we'd actually do
Anki for words and sentences, one folder in whatever notes app you already have for grammar and mistakes, and a phone capture habit fast enough to catch a phrase mid-conversation. Nothing else needed, and no reason to change notes apps for this.
If capture is the part that keeps failing — the phrase heard on the bus that never got written down — that is what Clair Mind is built for: one tap, offline, private, searchable later. It is not a flashcard app and will not schedule your reviews. Anki does that better, and it is free.
More roundups in best notes apps, and the study mechanics in memory and focus.