Frequently asked questions

Laesi is a desktop reading app for Mac, Windows, and Linux. You import a book, subtitle file, or web article; Laesi tokenizes and lemmatizes it; you click unknown words to look them up and set a knowledge level; over time the text transitions from mostly unknown (highlighted) to mostly known (clean).

Serious readers of any of 60+ languages — but especially the ones the big apps don't support or support poorly: Icelandic, Old Norse, Faroese, Finnish, the Sámi languages, Welsh, Irish, Kalaallisut, Latin. The mainstream European languages (Spanish, French, German, and the rest) are fully supported too.

Mac (macOS 12+, Apple Silicon), Windows (10 or 11, 64-bit), and Linux (.AppImage / .deb). One purchase covers all three — install on as many machines as you personally use. A couple of analyzers (Finnish via Voikko, the GiellaLT family) need an extra toolchain on Windows; Mac and Linux get the full set out of the box.

$99, one-time, with every language and every feature included. No tiers, no add-ons, no subscription. One license covers Mac, Windows, and Linux on the machines you personally use.

No. We offer a 30-day money-back guarantee instead — buy it, use it for a month, email us for a full refund if it isn't for you. No questions asked.

No. Laesi is a one-time purchase. No monthly fee, no yearly renewal, no servers you're paying to rent.

50% off with a .edu (or equivalent) email address — email us. Academic and institutional site licenses are also available for Latin, Old English, and classical-language programs.

Free updates for life. Buy Laesi once and every update is included — new features, new analyzers, bug fixes — for as long as we keep building it. No annual renewal, no paid 'Laesi 2.0', no upgrade fees.

60+ and growing — every one included in the single price. That spans the Nordic and North Germanic languages (Icelandic, Faroese, Old Norse, Norwegian, Danish, Swedish), West Germanic (German, Dutch, and minority varieties), Romance, Slavic, Baltic, Greek, Maltese, the Uralic family (Finnish, the Sámi languages, Kven, Meänkieli), Celtic (Irish, Welsh), Turkic, indigenous languages (Kalaallisut, Ojibwe), and a range of world languages. Latin, Old English, and Middle English are on the roadmap. See the full list.

Software that maps an inflected word form back to its dictionary entry. In Finnish, kirja (book), kirjan (book's), and kirjoissa (in the books) are all forms of one word. Without a lemmatizer, your vocabulary list balloons with duplicates and your progress tracking breaks. Laesi ships real linguistic lemmatizers — Voikko for Finnish, spaCy across the modern European languages, and GiellaLT for the Sámi family, Faroese, Irish, and Old Norse. Languages without a dedicated analyzer use Wiktionary surface-form lookup, and we keep wiring up new analyzers over time.

Each language's page shows its exact status. Full morphological lemmatization covers Finnish, the modern Germanic/Romance/Slavic/Baltic languages, Greek, Maltese, and the GiellaLT languages (Sámi, Faroese, Irish, Old Norse, Kalaallisut, Central Ojibwa). The rest — including Icelandic, Welsh, Czech, Hungarian, Turkish, and most world languages — use surface-form lookup today, with analyzers planned.

If there's an open-source morphological analyzer we can wire in (GiellaLT, CLTK, spaCy, Stanza, Hunspell) and a Wiktionary section for it, probably yes. You can also add it yourself right now: Laesi's custom-language tool handles minority dialects, conlangs, and historical varieties.

They need tokenization and layout strategies that differ from Laesi's current assumptions — right-to-left scripts (Arabic, Hebrew) and character-based segmentation (Chinese, Japanese). Support for RTL and CJK languages is on the roadmap, not available yet.

Yes — for Chrome and Firefox. It highlights words on any web page using your local vocabulary, lets you set a word's level inline, runs Wiktionary/Google lookups, and can translate the surrounding sentence. Bulk 'mark all seen/known' and a Record-Reading button mirror the desktop reader.

Two paste-based workflows: a Graded Reader that rewrites a chapter to your vocabulary or a CEFR level, and translate-and-annotate that builds an interlinear overlay. Laesi generates the prompt, you run it in whatever LLM you like, and paste the response back. There's no API key to configure, and Laesi itself doesn't send your text anywhere.

Yes — the Facsimile Reader renders scanned page images with tappable word blocks positioned over the original layout, so you can read and look up words on a facsimile.

In a SQLite file on your computer. Nowhere else. Default location: ~/.laesi/laesi.db. Back it up however you back up the rest of your files.

Only for dictionary lookups you trigger — the Wiktionary or Google Translate call sends just the word, not who you are. No telemetry, no analytics, no user tracking. The AI features run in your own LLM, not ours.

Put the SQLite database in a Dropbox / iCloud / OneDrive folder and point Laesi at it. There's no built-in cloud sync — we don't run a server. You can also export and re-import your profile.

As many as you personally use. Laesi's license is per-person, not per-device. Don't share your license key with friends; do put it on your laptop, desktop, and work machine.

Not yet — bulk vocabulary import is on the roadmap. Today you start from scratch, though reading fills your vocabulary fast, and you can move your own Laesi profile between machines with export and import.

Not yet. A companion mobile reader is on the roadmap. For e-ink, Boox (Android) is the realistic path when that lands.

It uses Wiktionary and Google Translate for dictionary lookups, so yes for those. Everything else — reading, lemmatization, highlighting, Anki export — works fully offline.