About

    About Capuchino

    Capuchino helps advanced Spanish learners (past B1) master the long tail of vocabulary that stands between them and C1 — the roughly 10,000 high-frequency words where most classroom curricula run out of road and most apps stop optimizing. It's aimed squarely at learners preparing for the Instituto Cervantes DELE C1 exam, but useful to anyone who has hit the upper-intermediate plateau.

    At its core, Capuchino is a personal gap-finder. A short temperature-check assessment estimates where your knowledge starts to fall off across frequency bands. From there, you refine your gap list by hand — “actually I know this,” “actually I don't” — and export a clean deck to the spaced-repetition tool you already use (Anki first; format-agnostic by design). We don't try to replace your SRS; we try to tell you exactly what to put in it.

    Capuchino is named after the capuchin monkey (“mono capuchino” in Spanish), found from Central America to northern Argentina.

    Data sources

    Capuchino's lexicon is ingested via a separate Python pipeline that reads a pinned snapshot of doozan/spanish_data — a compilation drawing on the sources below. Imported data lives in read-only tables (the app physically cannot write to them); our authored content (DELE tags, learner activity) lives separately so license obligations have a clean boundary.

    SourceUsed forLicense
    doozan/spanish_dataCompilation we ingest from: 10k lemmas + 21k senses + 108k example sentences + 577k formsCC-BY 4.0 (compilation)
    English WiktionaryEnglish glosses, qualifiers, synonyms, gender, etymology (via doozan)CC-BY-SA 4.0
    hermitdave/FrequencyWordsWord-frequency rankings, subtitle-derived (via doozan)CC-BY-SA 3.0
    TatoebaBilingual example sentences, filtered to quality ≥ 5 (via doozan). Per-sentence author attribution shown wherever sentences are displayed.CC-BY 2.0 FR

    Accessibility

    We aim to follow web accessibility best practices (WCAG 2.1 AA) so that Capuchino is usable by as many people as possible — including with screen readers, keyboard-only navigation, and reduced-motion preferences.

    Third-party tools

    This project is built on open-source software. Below are the main libraries and services we rely on.

    Core framework

    UI & design

    Data & forms

    Email

    • Resend

      Transactional email delivery

      Proprietary API

    Testing

    Data pipeline

    • Python

      Lexicon ingest scripts in a separate repo (run outside the app)

      PSF
    • uv

      Python package & version manager

      MIT/Apache-2.0
    • supabase-py

      Python client for Supabase (used with the service_role key)

      MIT
    • tqdm

      Progress bars for long-running ingest jobs

      MPL-2.0/MIT

    AI

    • Claude Code

      Used as an Obsidian community plugin, in combination with Obsidian, for planning, PRD work and architecture notes

      Proprietary

    Platform & hosting

    • Lovable

      AI-assisted app development platform

      Proprietary
    • GitHub

      Source code hosting & version control

      Proprietary