FAQ
Frequently asked
Real questions we've gotten — or ones the marketing copy doesn't fully answer. Short, direct, no marketing-speak.
Who is this for?
Is this only for English speakers learning Spanish, or also Spanish speakers learning English?
Capuchino is built for learners of Spanish — typically English speakers, but anyone is welcome. The mechanic itself isn't language-locked, and our multi-list schema could in theory support other target languages. But Spanish is the v1 focus and the only target language on the near roadmap. Spanish speakers learning English would need a different product.
Do I need to already be at B1 or B2 to use Capuchino?
Capuchino works best for learners past B1. The assessment will identify your "cliff" — the frequency band where your knowledge starts to drop off — and ours is calibrated for cliffs in the 4,000–7,000 range. If you're earlier than that, the gap list will be enormous and not particularly actionable. Take the assessment regardless; if your cliff lands at band 1 or 2, classroom curricula will serve you better for now.
The 10,000 number
Is 10,000 words really "C1"?
Roughly, with caveats. The top 10,000 high-frequency words give you ~97% lexical coverage of typical Spanish text — the threshold most research cites for C1 reading comfort. But the CEFR itself doesn't specify a vocabulary count, and academic estimates vary (most put C1 in the 7,000–10,000 range). The point of Capuchino isn't the absolute number; it's identifying *your* gap so you can close it.
Where does the 10,000 list come from?
Subtitle-derived frequency data via doozan/spanish_data — a compilation drawing on OpenSubtitles 2018 frequencies, Wiktionary glosses, and Tatoeba bilingual example sentences. Full credits and licenses on /about. Subtitle data has a known bias toward conversational register; we don't pretend otherwise.
How it works
How long does the assessment take?
About 10 minutes for 50 items, stratified across the ten frequency bands. You can pause and resume mid-flight — your answers are saved as you go.
Why self-rating instead of multiple-choice?
You know what you know better than a four-option test does. Self-rating (solid / shaky / unknown) on a real example sentence is faster, less brittle, and more honest than guessing-by-elimination on multiple choice. The trade-off: it requires intellectual honesty. Try not to lie to yourself.
Why feed Anki instead of building your own SRS?
Spaced repetition is solved. Anki, Mochi, RemNote and others all do it well, and most advanced learners already have a daily SRS habit. We feed that habit with a clean export of *your* gap; we don't ask you to switch tools or build a new one. If you don't use any SRS yet, Anki is free and well-documented — start there.
What's not here yet
What's not in Capuchino yet?
A lot. No audio yet — neither TTS nor human. Only one frequency list (doozan-derived; additional sources are gated on licensing). Self-rating is basic — no productive recall, no fill-in-the-blank training. One topic published so far (the human body); seven more are coming. No social layer yet. See /log for what's shipped.