Look up a word without leaving the sentence. What you save is filed under the book it came from and comes back until you know it — a deck made of your own reading, not somebody else’s vocabulary list.
Nothing asks you what to study. The book decided that when you chose it, and its words wait under its name until they are yours.
Select any passage for its grammatical structure, colour-coded parts of speech, and an in-context translation beside the original. The reason a line was hard is usually its shape, not its vocabulary.
Save passages with notes attached. Review saved words on flip-style cards. Export the best of either as a share card.
Lucerna paginates them, builds a table of contents, and indexes the full text for instant search — ready in minutes. Chapters wrong? Edit the table of contents yourself.
Public-domain books in 12 languages, ready to open — same lookup, same account, same review.
Not the light theme inverted: ink-navy stock, bone type, brass marks. Two surfaces, and the paper tint and texture are yours to set on either. Light from 06:00 to 18:00 and 夜航 in the evening, by your local clock — or pick one and stay there.
Upload a book in any of the seventy-one languages the reader recognises and it opens as itself — paginated, searchable, its chapters found. How much more you get depends on the language: full sentence analysis in three, word lookup and translation in five more, reading and vocabulary in the three CJK scripts.
What comes back — definitions, translations — is in the language you already speak: it follows the interface unless you set it separately. Analysis depth tracks the model you point Lucerna at, so a stronger model moves a language up.
It runs translation and analysis on local models — Ollama, LM Studio, llama.cpp. Offline, no key, no bill, books kept on your device, no account limits.
Why do we read? Why learn a foreign language? For a long time I've felt they are the same question — both take you out of your own world, and let you come back carrying something unfamiliar. This is why we're building Lucerna.
The honest numbers: about 5,000 word families for rough comprehension, 8,000-9,000 to read comfortably on your own. But the more useful answer is that you don't have to wait until you get there — reading is how you get there.
The hardest part of your first novel in a new language isn't the grammar — it's not knowing what happens next. So take that problem off the table: read a book you've already read.
Meet the language again, and again.
Pick a book you actually want to finish. The rest follows from it.
Reading lists, language tips, and a quiet place to share what you’re working through. Pick the platform you live on.
Lucerna is a multilingual ebook reader for learning languages through real books. Tap a word for its definition, pronunciation, and conjugation; select a sentence for a grammatical breakdown and translation; save words to a review deck — all without leaving the page.
Yes. You can try it without an account (1 book, 5 saved words), and a free account raises those limits (5 books, 200 saved words) and syncs across devices. AI translation and sentence analysis run on your own LLM API key (OpenAI, Anthropic, or Google) — their free tiers are usually enough for reading. Or skip the cloud entirely: the desktop app works with local models (Ollama, LM Studio, llama.cpp…), runs offline, costs nothing, and keeps books on your device — no account limits.
The reader works with 11 languages: English, German, French, Spanish, Italian, Portuguese, Dutch, Russian, Japanese, Korean, and Chinese. AI translation and sentence analysis currently cover English, German, and French, with more on the way.
Yes — upload EPUB, PDF, or TXT files. Lucerna paginates them, builds a table of contents, and indexes the full text for instant search, so your own books get the same lookup and analysis as the built-in library.
No. Try Mode works without one. When you want to keep more books and vocabulary — and sync them with the desktop app — a free account takes a minute to create.
You never leave the sentence you are reading. Looking up a word or translating a passage happens in place, and every word you save lands in a review deck — instead of switching apps and losing the thread.
Yes — macOS, Windows, and Linux, with the same account and library. The desktop app adds offline reading, faster text-to-speech, and OS-native shortcuts.