🇱🇺 LuxFlash

Learn Luxembourgish (Lëtzebuergesch) with flashcards — free, open-source, in your browser.

▶ Open the app

LuxFlash is a no-account, no-tracking flashcard web app for learning the most common Luxembourgish words. It runs entirely in your browser: 2000+ words across CEFR levels A1 → B2, each with an example sentence, German & French glosses, and a direct link to the official Lëtzebuerger Online Dictionnaire (lod.lu) for audio, gender and full declensions.

Your progress is saved locally (browser localStorage, flushed on tab close and protected with persistent storage where the browser allows it) and you can download or upload a JSON backup to move between devices.

---

✨ Features

• Spaced repetition — a lightweight Leitner system schedules each card so you review words right before you'd forget them (Again / Hard / Good / Easy).
• Levels — filter your study deck by A1, A2, B1, B2 (mix and match).
• 2000+ curated words — organised by theme (family, food, work, travel, nature, verbs, grammar words, adjectives…), each with a natural example sentence and its translation.
• Reverse mode — practise English → Luxembourgish recall.
• Browse & search — scan the whole dictionary, filter by level/category.
• Grammar tab — concise, practical notes (gender & articles, cases, plurals, verb tenses, word order, the Eifeler Regel, numbers, questions…).
• Resources tab — curated links to keep learning (Luxembourgish with Anne, LOD, RTL, ZLS and more). LuxFlash links to these, it doesn't reproduce them.
• Native pronunciation — 1800+ words play the real native recording from lod.lu (open CC0 audio, verified against the official LOD open-data dump), with optional auto-play and a browser-speech fallback for the rest.
• lod.lu integration — every card and list entry deep-links to LOD so you can verify spelling, hear the pronunciation and see the full grammar.
• Progress dashboard — words learned, due today, day-streak, per-level bars, Leitner mastery chart.
• 100% local & private — nothing is uploaded. Export/import a JSON backup any time.
• Pronunciation — uses the browser's speech engine (best-effort) to read words aloud.
• Dark mode, fully responsive, keyboard shortcuts, no build step, no dependencies.

⌨️ Keyboard shortcuts

Key	Action
Space	Flip card
1 2 3 4	Grade: Again / Hard / Good / Easy
P	Play the pronunciation

🚀 Run it locally

It's a fully static site — just open it.

git clone https://github.com/alexff91/luxflash.git
cd luxflash
# either open index.html directly, or serve it:
python3 -m http.server 8000   # then visit http://localhost:8000

🌐 Deploy on GitHub Pages

This repo ships a GitHub Actions workflow (.github/workflows/pages.yml).

1. Push to your main branch.
2. In the repo: Settings → Pages → Build and deployment → Source: GitHub Actions.
3. The site publishes to https://<user>.github.io/luxflash/.

(You can also just set Source: Deploy from a branch → main → / (root).)

🗂️ Project structure

index.html              app shell
assets/css/styles.css   styling (light/dark)
assets/js/storage.js    localStorage + JSON import/export
assets/js/srs.js        Leitner spaced-repetition logic
assets/js/app.js        UI controller
data/words.js           the dataset (generated) → window.LUXFLASH_WORDS
data/words.json         same data as plain JSON
data/chunks/*.json      themed source batches
scripts/build_data.py   merges + de-dupes chunks into data/words.*

🧱 Editing / extending the vocabulary

Each entry looks like:

{
  "lb": "Haus", "en": "house", "de": "Haus", "fr": "maison",
  "pos": "noun", "gender": "n", "level": "A1", "category": "home",
  "ex_lb": "D'Haus ass grouss.", "ex_en": "The house is big."
}

Add or fix entries in any data/chunks/chunkN.json, then rebuild:

python3 scripts/build_data.py

This regenerates data/words.js / data/words.json (validated, de-duplicated, sorted by level). Commit the regenerated files.

🔊 Native LOD audio

LOD publishes pronunciation audio openly (CC0) at https://lod.lu/uploads/OGG/<id>.ogg and …/AAC/<id>.m4a, where <id> is the LOD entry id. To wire real recordings to LuxFlash words, run (on a machine that can reach lod.lu / data.public.lu):

python3 scripts/fetch_lod_audio.py --verify   # match words + HEAD-check audio, writes "lodId"
python3 scripts/build_data.py                 # rebuild data/words.js

The script auto-discovers the newest LOD dump via the data.public.lu API, matches headwords part-of-speech-aware, and (with --verify) keeps only ids whose audio file actually exists. Use --local file.zip to reuse a downloaded dump and --force to re-match existing ids.

The 🔊 button then plays the native recording, and falls back to the browser's speech engine for any word without a match. The per-word LOD ↗ link always opens the official entry with audio regardless.

🙏 Vocabulary accuracy & lod.lu

The vocabulary is community-maintained and was seeded automatically. It is a learning aid, not an authority — for definitive spelling, grammatical gender, plurals, pronunciation and declensions, always consult the official Lëtzebuerger Online Dictionnaire — lod.lu, which every card links to directly. Spotted a mistake? Please open an issue or pull request — corrections are very welcome. 🐛

📄 License

MIT. LuxFlash is not affiliated with or endorsed by lod.lu / the Luxembourgish government; it only links to their public dictionary.