Why do some accent removers leave ø and ł unchanged?

Those letters have no Unicode decomposition, so a normalize-and-delete-combining-marks method never sees a removable accent. Remove Accents from Text applies an explicit fold table for them, verified against the Unicode Character Database.

Will it damage Greek, Cyrillic, or Asian text?

No. Marks are removed only from Latin base letters. Greek, Cyrillic, Chinese, Japanese, Korean, Devanagari, and emoji pass through byte-for-byte by default; enabling strict mode deletes non-Latin characters rather than romanizing them.

Is turning ß into ss really removing an accent?

No. It is a transliteration because one letter becomes two. The tool presents it as a separate choice, enables it by default for data cleaning, and lets you switch it off to keep ß intact.

Can I run the cleaned output through the tool again?

Yes. The operation is idempotent, so processing an already-cleaned result again makes no changes. This keeps repeated pipeline use predictable.

Remove All Accents From Text: Control Every Change

To remove all accents from text, paste it into the browser-based cleaner, choose the relevant Latin-letter options, click Remove accents, and copy the cleaned result. This Remove Accents from Text tool turns café into cafe, señor into senor, and Zürich into Zurich by removing diacritical marks from Latin base letters. It also handles letters that ordinary accent stripping can miss: ø becomes o, ł becomes l, and đ becomes d, so Łódź becomes Lodz rather than remaining partly unchanged. Separate controls let you spell out transliterations such as æ to ae, œ to oe, ß to ss, and þ to th, and decide whether non-Latin scripts should remain in the text. Greek, Cyrillic, Chinese, Japanese, Korean, Devanagari, and emoji pass through unchanged under the default preservation settings, while Latin diacritics are removed only where intended. Processing happens in your browser, so the pasted text is not uploaded, stored, or attached to an account. For mixed-language data, keep the default non-Latin preservation setting and review the separate transliteration choice before copying the result. The page reports exactly how many characters changed, giving you a quick check before the cleaned text enters a matching or import workflow.

Know what gets changed

Accent removal is narrower than replacing every non-ASCII character. A diacritic is a mark attached to a Latin base letter, while several useful Latin letters have no separable mark to delete. That difference explains why a basic Unicode normalization routine can turn é into e but leave ø, ł, or đ unchanged.

The Unicode Character Database records no decomposition for that second group. The cleaner supplements normalization with an explicit fold table checked against the database. This is why ø, ł, and đ receive their base-letter forms even though normalization alone cannot produce them, and why Łódź becomes Lodz.

Input class	Examples	Default result	Type of change
Latin text with diacritics	café, señor, Zürich	cafe, senor, Zurich	Diacritical marks are removed from Latin base letters.
Non-decomposing Latin letters	ø, ł, đ	o, l, d	Explicit Latin folds are applied.
Documented transliterations	æ, œ, ß, þ	ae, oe, ss, th	The separate transliteration toggle is on by default.
Protected non-Latin content	ά, ё, emoji	Unchanged	Non-Latin preservation is on by default.

These categories should not be blurred. Stripping a diacritic, spelling one character as two, and deleting an entire non-Latin character are different operations. Keeping them separate lets you choose an output that matches the job instead of treating every changed character as an accent.

Set the cleanup behavior

The transliteration toggle is on by default because data-cleaning workflows commonly want ß, þ, and ligatures represented as ordinary letter sequences. When the toggle is on, documented mappings such as æ to ae and œ to oe are applied along with accent removal. Turn it off when the exact one-letter forms must remain, such as in a linguistic archive, while still stripping true Latin diacritics and folding non-decomposing Latin letters.

The non-Latin setting answers a different question: should scripts outside the Latin alphabet be preserved or removed? Preservation is the default. Optional strict mode is off by default; when enabled, it deletes non-Latin characters rather than transliterating them. Use it when the destination truly requires those characters to be removed for ASCII-only output.

General data cleaning: Keep transliteration on and strict mode off to use the documented default behavior.
One-letter spelling preservation: Turn transliteration off to retain letters such as ß and þ, and keep strict mode off to preserve other scripts.
ASCII-only import: Enable strict mode when non-Latin content must be deleted, while checking that permanent deletion fits the workflow.

Clean the text in four actions

Paste the source text. Add the accented names, special Latin letters, or mixed-language content you want to clean. The input limit is one million characters per paste.
Choose the transliteration and script settings. Leave transliteration on for the default data-cleaning output, or turn it off to retain one-letter spellings. Keep non-Latin scripts under the default preservation setting, or enable strict mode when they must be deleted.
Click Remove accents. The operation processes the paste in one linear pass, removing marks from Latin base letters, applying explicit folds, and following the selected transliteration policy.
Check the change count and copy the result. Review the reported number of changed characters, spot-check special letters, and then copy the cleaned text into the next task.

Protect mixed-language content

The removal scope is deliberately limited to marks attached to Latin base letters. Greek and Cyrillic can decompose in a similar way to accented Latin, but the default settings do not strip their marks. Devanagari vowel signs are treated as grammatical letters rather than removable accents. Chinese, Japanese, Korean, Devanagari, and emoji also pass through byte-for-byte when non-Latin preservation is enabled.

Precomposed and decomposed input produce identical results. A source can store é as one Unicode code point or as a base letter followed by a combining accent, and the cleaner returns the same normalized result in either case. It recomposes the output to NFC, the precomposed normalization form, so later systems receive a consistent representation.

The operation is also idempotent: running an already-cleaned result through it again changes nothing. That property makes the output predictable in repeated pipelines. Before copying a large batch, inspect an ordinary accented word, a non-decomposing Latin letter, and any nearby emoji or non-Latin text. The change count shows how many characters changed, while the spot-check helps confirm that the selected policy fits the data.

Use the output in a data workflow

Treat the cleaned string as a normalized value for comparison, matching, or import. It does not need to replace the original display text, and additional formatting rules may still be required for the destination system.

Deduplicating a contact list: Create an accent-free comparison key while retaining the original accented name in a separate display field. This can reveal records that differ only by diacritics without discarding their original spelling.
Matching names across systems: Clean both sides with the same transliteration policy. Decide explicitly whether mappings such as ø to o and ß to ss meet your matching rules.
Building identifiers or slugs: Use the accent-free result as one normalization stage, then apply the case, whitespace, and separator rules required by the destination.
Preparing ASCII-only software: Enable strict mode when non-Latin characters must disappear, while understanding that they will be deleted rather than converted into Latin letters.

Several distinct spellings can produce the same cleaned output, so review the mapping policy before using that output as permanent display text. Keeping the source value and the normalized value separately makes later corrections and audits easier.

Understand the tool’s boundary

This is a Latin-script diacritic remover with documented transliterations, not a general transliterator. It will not romanize Russian, produce pinyin from Chinese, or convert Greek into Latin letters. Strict mode can delete non-Latin characters, but it does not translate them.

Specialist letters such as ð and ĸ are folded by convention rather than by a Unicode decomposition rule. The page discloses those conventional mappings instead of presenting them as automatic accent removal. Other explicit fold-table entries are checked against the Unicode Character Database.

For mixed-language text, the default combination is transliteration on and strict mode off. For true diacritic-only cleanup, turn transliteration off. For an ASCII-only destination, use strict mode only when deleting non-Latin content is intentional, then inspect the cleaned result before import.

If you're weighing options, How to Remove Duplicate Words in Google Docs covers this in detail.

If you're weighing options, How Do I Avoid Mistakes When I Remove Punctuation? covers this in detail.