Accent and diacritic remover

Accent remover

What the accent remover does

This app strips the accent marks off letters that carry them: café turns into cafe, naïve into naive, jalapeño into jalapeno. Letters, words, punctuation and line breaks all stay exactly where they were. It's a one-click pass for the cases where an accented character causes a problem somewhere downstream.

It's local. Everything runs in your browser, nothing you paste is uploaded, and there's no cap on how much you clean or how often. That matters here more than on most tools, because the usual reason to strip accents is generating a database key or a filename out of a real person's name.

Workflow

How to remove accents from text in three steps

Nothing to configure.

Step 1

Paste text with accented letters into the box.

Step 2

Press Remove accents. It runs instantly, on your device.

Step 3

Copy the plain-letter result.

The mechanism

Decompose then strip: how NFD makes this work

An accented letter like é can be stored in Unicode two ways: as one precomposed character, or as a plain e followed by a separate combining acute accent, two codepoints that render together as one glyph.

This app applies NFD, canonical decomposition, which forces every letter into that second form. It then deletes everything in the Unicode Mark category, which is exactly the set of combining accents, and recomposes the remainder with NFC. That's why the base letter always survives clean: decomposition separates it from its mark completely, so deleting the mark never touches the letter underneath.

Coverage

Which scripts and languages this works on

Because it removes marks by Unicode category rather than from a Latin lookup table, it doesn't stop at the edge of the Latin alphabet the way most accent removers do.

Latin

French, Spanish, Portuguese, Italian, German, Polish, Czech, Turkish, Romanian and the Nordic languages. Acute, grave, circumflex, tilde, umlaut, cedilla, caron, macron and ring all come off.

Greek

Tonos and dialytika are combining marks, so Ελλάδα comes back as Ελλαδα.

Cyrillic

Й becomes И and ё becomes е, since the breve and diaeresis decompose exactly like a Latin accent.

Arabic and Hebrew

Harakat and niqqud are stripped, leaving the consonant skeleton. Useful for normalising search, not for publishing.

Vietnamese

Both the vowel mark and the tone mark come off, so Tiếng Việt becomes Tieng Viet. That discards tone, which is meaning, so keep it to technical fields.

A real limit

What this app will not convert

This only works on characters that decompose into a base letter plus a separate mark. Some letters look like accented cousins of a plain letter but are distinct letters in their own alphabet, with nothing to decompose. Competing pages list several of these as converted. This one doesn't, because it doesn't.

Ø and ø

Danish and Norwegian o with stroke. The stroke is part of the letter, so it stays.

Ł and ł

Polish l with stroke. Unchanged.

Đ, Ð and đ

Croatian, Icelandic and Vietnamese d with stroke. Unchanged, even in Vietnamese text where every other mark comes off.

Æ, æ, Œ and œ

Ligatures rather than accented vowels. Nothing decomposes, so nothing is removed, and they don't become ae or oe here.

ß

The German sharp s is a letter, not an accent. It doesn't become ss.

Å and å

The surprise: the ring above is a combining mark, so this one does reduce to a, unlike its neighbour ø.

Why it matters

Why café becomes cafe: what this is for

Plenty of older systems, some slug generators, certain legacy database fields and a range of search setups either can't handle accented characters or handle them inconsistently, so the same word entered two ways fails to match itself. Stripping accents produces a version that works anywhere plain ASCII letters are expected: slugs, filenames, usernames, email local parts, programming identifiers and legacy imports.

It's also useful purely for matching. Someone searching cafe should still find café. Indexing both the original and the stripped version lets a plain-letter query catch either one, which is how most search systems handle this internally anyway.

What accents cause

What accented characters break downstream

A slug generator that mishandles é produces a broken-looking link or drops the character instead of transliterating it. Sorting a name list under weak Unicode collation places é nowhere near e. Older filesystems and single-byte exports mangle accented characters into question marks or garbled bytes you can't recover.

In spreadsheets it shows up as a lookup that fails against a value that looks identical on screen, because one side holds a precomposed é and the other holds e plus a combining accent. Run both columns through here and they agree.

Who it's for

Who uses an accent remover and when

Developers generating URL slugs, filenames, usernames or database keys from user-entered text, where the source language is unpredictable and the target system isn't. Anyone exporting into a legacy system with limited character support runs the data through here first.

It's also standard in search and matching work: building an index that finds café whether someone typed café or cafe, or reconciling name lists from systems that disagree about accents.