Remove diacritics
Remove diacritics. Runs entirely on your device, nothing is uploaded.
Runs on your device. The file is never uploaded.
Remove diacritics decomposes a text file, drops every character Unicode files under Mark, and recomposes, so an accented Latin letter arrives bare. No table and no settings take part. A letter that is one character rather than a base plus a mark survives whole: the sharp s, the slashed o, thorn and the barred l.
Questions
What does it do to my text exactly?
It normalises to NFD, deletes every character in the Unicode Mark category, and normalises back to NFC. So e with an acute accent decomposes into a plain e plus a combining acute, the combining acute is removed, and you are left with e. That covers the accents, cedillas, umlauts, carons and rings that sit on top of a Latin base letter, in one pass, with no lookup table.
Why did the German eszett and the Nordic letters survive?
Because they are not a base letter plus a mark. Characters like the sharp s, ae, oe, the slashed o, eth, thorn and the barred l are single letters in their own right; NFD has nothing to decompose, so nothing is stripped. Use the Transliterate tool instead, which carries a small table mapping those to ss, ae, oe, o, d, th and l.
Is this the same as transliteration?
No. This tool only takes marks off Latin letters. Cyrillic, Greek, Hebrew, Arabic and CJK text passes through completely unchanged, because those scripts are not accented Latin. Transliterate goes further: it does this same decompose-and-strip pass and then applies a table for basic Cyrillic, basic Greek and about eight extra Latin letters. Pick this one when you want the original script kept and only the accents gone.
Will it break Japanese, Korean or Indic text?
Japanese and Korean are safe in practice, but be careful with scripts where marks carry meaning. The filter deletes every Unicode Mark, and in Devanagari, Thai, Arabic and Hebrew vowel points and matras are Marks. Those characters will be removed and the text will change meaning. The tool is built for Latin script with accents; check the output before using it on anything else.
Are there any settings?
No. There is nothing to configure, the transformation is fixed, and the same input always gives the same output. It runs per file, so you can drop a batch and get one cleaned result per file, all of it inside a Web Worker in this tab with no upload.