Beige flashcard grind.
Digital immersion, done properly.
Spaced-repetition cards your AI builds for you, with the voices already generated, so you can focus on actually drilling them.
Your cards, in action.
Five kinds of card — meaning, word, listening, cloze, translation — and each one comes back on its own schedule. Pick a language; the whole page follows.
the arrows steer, or let it run · more languages coming soon
Multidirectional cards.
One word in, four cards out. Recall has directions, and each direction is its own memory — so every entry is tested from every side, each on its own schedule.
four memories · four schedules · one entry
Every card arrives with its voice.
The agent writes the pronunciation, not just the text — stress marks and tones go in before the voice is made, and a second pass checks them. (If one ever comes out wrong, tell it to try again.)
this is the pipeline, unedited · press one
Bring everything over from Anki. Nothing lost.
Drop in your export and every deck arrives with its history intact: due dates, intervals, lapses, review log, audio and images. Tomorrow's reviews are tomorrow's reviews — you pick up exactly where you left off.
reads any .apkg · your schedule carries over
Decks built in bulk, not card by card.
One sentence in, fifty cards out — nouns, verbs, phrases, whole sentences — each arriving with its audio and its picture already made.
you can still build cards one at a time, the way you always have
Curated decks to get you started.
Pre-made decks in every language, ready on day one — starting with the sounds your ear was never trained to split.
Minimal-pair decks, verified by native speakers
Pairs a native speaker hears as two different words can reach your ear as one — your language never needed the split. Play one, prove which you heard.
every language · more decks coming soon
Practice conversations, built from your cards.
Ask for one — about the dog, about the mountain, about whatever your cards can carry — and your AI writes a short conversation using only words you know, with the voices already made. Play it, read along, tap a line to hear it again.
beta · phrases and sentences, never a bare word · the pictures are your cards’ own
made from these cards
Connect your AI. Ask anything about your cards.
One sign-in inside the app and any Claude conversation can build, fix and tend your decks. ChatGPT and Gemini work by copy and paste.
no connector needed: copy your cards out, paste them into any chat, hand the answers back
Ask where you go wrong.
It reads your review history — every lapse, every slow answer — and names what is actually tripping you up. Then it builds the cards for exactly that.
reads your review history · answers in the chat you already use
Flag cards as you go.
A tap in the corner, and you keep reviewing. A flagged card just carries a tag — it waits for you and your AI to look over later, to change it or to learn from it.
wrong voice, weak example, a question you want answered: tag it and move on
Then hand them over.
“Let’s go through the cards I flagged.” Your AI answers your questions and makes the changes you asked for — a new voice, a better example, one card split into two — and holds every change for your approval.
or sit down with the flagged pile yourself, it is just a filter
Why cards work.
Your knowledge is a network, not a list. A new word hops in through what is already there — the picture you know, the English word — and every review grows it pathways of its own, until it fires the pattern by itself with no English in the loop.
None of your time is wasted.
Two findings have held for decades: you acquire a language from input you almost understand, and you keep what returns on a widening schedule. The scheduler is built on both, so every review lands at the moment it counts.
Both grew into the same modern scheduler, which returns every card at the last moment it still counts.
Catch a word just before it slips and it rebounds higher — stronger memory, slower decay, longer gap. The scheduler finds that edge for every card; ten minutes a day holds a vocabulary of thousands.
Everything in the box.
You will look at this a few hundred times a month. Hover anything.
local-first · your cards never leave your machine
Ask in one sentence; an agent writes the cards, files them, and records the audio.
Neural voices tuned per language get stress, palatalization, and tone right — no hunting mp3 files, ever.
A memory model trained on billions of reviews — each card returns at the exact moment before you would forget it. The closest thing to downloading vocabulary into long-term storage.
Pairs a native speaker hears as different can sound identical to you — the trainer plays one and your ear proves which it heard.
Recall, meaning recall, cloze, relearn — the same word is tested from every direction until it is yours.
Talk with AI that stays inside the words you know — understanding everything you hear is what makes conversation practice work.
The AI reads your review history, finds where you are falling behind, and builds the lesson that catches you up.
A conversation to practise, written from the words you already know, new whenever you ask.
betaWhy immersion, and why now.
Immersion works, and you already live online. Your feed, your podcasts, your AI — all of it can happen in the language you are learning. The cards are the part of that habitat that makes the words stick.
Copy the list of words you know, paste it into any chat, and have spoken conversations that stay inside them.
Every podcast you already listen to ships audio tracks in other languages.
placeholder · founder video