Remove accents
Turn accented text into plain letters, without pretending it is lossless.
Accents come off, and the letters that are not accented forms of anything — ß, ø, æ, ł, þ — are written the way their languages write them in ASCII: ss, o, ae, l, th. München becomes Munchen. Choose ü → ue for the German convention used in addresses and domain names. This is a Latin-script operation: Cyrillic, Greek, and other alphabets are left as they are. Up to 200,000 characters.
Result
Start typing and the result appears here. No button needed.
Remove accents, without the extra steps.
Strip the accents off Latin text: café becomes cafe, Łódź becomes Lodz. Letters that are not accented forms of anything — ß, ø, æ, þ — are written the way their own languages write them in plain letters, and umlauts can follow the German convention where ü becomes ue. Anything outside the Latin alphabet is left alone, and the page says so.
How to use this tool
- 1Paste the text with accents in it.
- 2Choose whether to strip marks only or also rewrite ß, ø, æ and ł, and which umlaut convention to use.
- 3Read the result, then select Copy result.
When Remove accents is the right tool
- A legacy system rejects anything outside plain ASCII and a list of names has to go into it tonight.
- A filename or a URL fragment needs the accents gone before it travels through tools that mangle them.
- A German address list has to use the ue convention, not bare u, because that is what the postal data expects.
- A search index cannot match café against cafe, and both spellings need a normalised version.
Not every accented-looking letter is an accented letter
é is e with a mark on it and Unicode can take the mark off. ß, ø, æ, þ, ł and đ are not: they are letters in their own right, with no decomposition and no plain-Latin twin. A tool that strips marks leaves them untouched — which is correct but rarely what someone converting to ASCII wants — so rewriting them as ss, o, ae, th and l is a separate, explicit choice. Each of those spellings is the one its own language uses when the letter is unavailable, not an invention.
Case follows the word, not the letter
A single uppercase letter becoming two ASCII letters has to decide between Thor and THOR. The rule here is the surrounding word: Þór becomes Thor because what follows is lowercase, and ÞÓR becomes THOR because it is not. The same applies to the German convention — Äpfel becomes Aepfel while ÄPFEL becomes AEPFEL. Getting this wrong produces THor, which is the kind of output that makes a conversion obviously machine-made.
This loses information on purpose
Polish ł and l are different letters; so are Turkish ı and i, and Vietnamese tone marks distinguish entirely different words. Removing accents makes text that a legacy system can store, and it makes two spellings comparable for search — it does not make an equivalent text. Keep the original as the source of truth and treat the stripped version as a key or a fallback. The workspace reports how many letters were rewritten as more than one, and how many were left alone because they are not Latin at all.
A few quick answers.
Because it is not an accented s. German writes it as ss when the letter is unavailable, which is what this produces — and what a German reader expects. The same reasoning applies to æ as ae, ø as o, þ as th and ł as l: these are letters in their own right with their own conventional spellings, which is why they are a separate choice from stripping marks.
Both are used; German addresses, domain names, and official transliteration use Muenchen, while informal ASCII often just drops the dots. The umlaut control picks between them, and the choice only affects ä, ö and ü — ß is ss either way.
Nothing, deliberately. Accent removal is defined over the Latin alphabet; converting another script into Latin letters is transliteration, which needs language-specific rules and often more than one right answer. The workspace counts the non-Latin letters it left alone so the result is never silently unchanged.
No. Once é is e there is nothing left to say which e it was, and ss could have been ss or ß. That is why the tool is explicit about what it changes: keep the original text, and use the stripped version only where the original cannot go.
Yes, as one half of the job: normalising both the query and the indexed text with the same settings makes café match cafe. It has to be the same settings on both sides, though — one side with ü → ue and the other with ü → u will not match at all.
It is the accent-removing part of it. Text to slug also lowercases, replaces punctuation with a separator, and collapses the result into one path segment. Use this when you want the words themselves in plain letters, and the slug tool when you want a URL.