Press kit
A description you can paste, the figures behind the reform, and the logo, screenshots and demo animation as files. Everything on this page may be quoted or reproduced verbatim, without asking.
Euspell is a reform of English spelling built as an engineering project rather than a manifesto. Where past reforms handed people tables of rules to apply by hand, Euspell compiles the reform into a program: a 205000-word lexicon in which every word carries its parts of speech and its reformed spelling, plus a classifier that reads the surrounding words to settle the hard cases — so records the noun and recordz the verb come out right automatically. It is deliberately conservative: no diacritics, no new letters, most words untouched, and each word's silhouette preserved (night → niht). One engine drives a browser extension, the Eupub e-reader, and word-processor add-ins, plus machine-readable lexicons for speech synthesis and grammar checking. It is open source, and every conversion happens on the user's own device.
The figures
Read from the shipping lexicon at version 0.3.4, not from an earlier draft. Each is quotable as it stands.
| Lexicon entries | 205505, each carrying its parts of speech |
| Spellings that change | 41438 — about 1 in 5 |
| Left exactly as they are | 164067, almost precisely four fifths |
| Words needing context to choose | 5920 |
| Noun/verb -s disambiguation | 94% accurate, linear SVM over context |
| Pronunciation lexicon | 35000+ words with IPA, for speech synthesis |
| Part-of-speech lexicon | 46000+ entries, Penn Treebank tags |
| Tagset | CLAWS-7 — 138 tags, University of Lancaster |
| Reform revision | r1, versioned independently of the software |
| Licence | Programs GPL-3.0-or-later; lexicon data CC BY-SA 4.0 |
| Price | Free, with no account, no server and no telemetry |
Logo and animation
The mark is ჱ — Unicode U+10F1, the Georgian letter archaic he — in blue, #0000FF.
142 common words in euspell, morphing to traditional spelling and back. Silent, loops, and opens on euspell so a mail client that freezes it on the first frame still shows the reform.
Outlined paths with no embedded text, so it needs no font and scales to any size.
Square, transparent background, for placing on any ground.
Flat white-on-blue, for avatars and anywhere transparency is unwanted.
Screenshots
Captured from a real browser at the size a store listing asks for, which is also the size an article wants.
A reader has rewritten a passage by hand and been marked against the engine. Before, after and the reason in one frame — the strongest single image here.
The popup on a page containing records, showing the noun and the verb resolved differently in the same sentence.
The built-in viewer rendering a PDF in euspell without altering the file.
What a reader can turn on, off, and exempt.
The store tile, at the size a listing asks for.
The white paper
Spelling Reform: An Engineering Approach is the primary source for everything above — the argument, the encoding scheme, and the method behind each figure. It reads in the browser or downloads as a PDF from the paper page. The software it describes is all downloadable from one page, and the source is public in three repositories: ossiak/euspell for the engine, lexicon, extension and add-ins, ossiak/eupub for the reader, and ossiak/euspell-game for the conversion game.
AI assistance
The reform architecture and the main lexicon are the author's own. Anthropic Claude Opus 4.8/5 and Fable 5 assisted with software development, revisions to sections of the white paper, installation instructions, the design of this website, and video production. The paper carries the same disclosure in full, before its references.
Inside the products, machine learning does one narrow job: deciding which of two or more existing spellings applies in a given context — the records against recordz problem. Nothing is generated while you read. Every tool ships with a fixed lexicon, and the conversion that runs on your device is a lookup and a classifier, not a language model.
Contact
Kamran Ossia created Euspell and answers press directly: kamran@euspell.org. Interviews, a longer biography, and anything not on this page are all available by asking.
EUSPELL is a trademark of Kamran Ossia; United States registration is pending. The mark identifies this project and its software. Using the word for the spelling itself — describing text as being in euspell, or writing about the reform — needs no permission and is not restricted.
Kamran Ossia created Euspell. He received his Bachelor's degree from Sharif University of Technology and his Master's and PhD from the University of Toronto, all in Electrical Engineering, specializing in Control Systems, and holds a second Master's degree in Applied Mathematics from California State University, Fullerton. He worked as a software engineer in the nuclear power and medical device industries, and is now retired.