UTF-8 to ASCII

ASCII has only 128 characters: English letters, digits, basic punctuation and a few control codes. Converting UTF-8 text to ASCII means replacing everything else with the closest plain character, a method usually called transliteration.

1. Paste your text, or open a text or CSV file.
2. Keep Swap é → e on to turn accented letters, curly quotes, dashes and symbols into their ASCII look-alikes.
3. Choose what happens to characters that have no look-alike: a ?, nothing, or a code that keeps the character recoverable.
4. Copy the result or download it as a .txt file. The status line lists anything that couldn’t be converted.

Chinese, Japanese, Arabic and other scripts, emoji, and symbols such as ° have no ASCII look-alike. Pick what suits the system you are sending the text to. For Tōkyō 東京:

Only the last two keep the original recoverable. Once a character has become ? or been removed, nothing can tell you what it was, so keep the UTF-8 original.

Not quite: UTF-8 contains ASCII. The first 128 UTF-8 characters are exactly ASCII, stored as the same single bytes, so a plain ASCII file is already a valid UTF-8 file. UTF-8 then adds over a million more characters, each stored as 2 to 4 bytes. You can see those bytes with the UTF-8 Converter.

You only need to convert when something accepts ASCII alone: older databases and mainframes, some bank payment references, label printers, legacy CSV imports, or file names and user names that must stay simple. Everywhere else, keep UTF-8, since converting always loses something.

# Command line (Linux)
iconv -f UTF-8 -t ASCII//TRANSLIT input.txt > output.txt

# Python: pip install unidecode
from unidecode import unidecode
unidecode("Zoë’s Café — €5")        # "Zoe's Cafe -- EUR5"

// JavaScript: accents only
text.normalize("NFKD").replace(/[\u0300-\u036f]/g, "")

// PHP (intl extension)
transliterator_transliterate("Any-Latin; Latin-ASCII", $text);

Each one behaves a little differently. iconv’s results depend on the system: on Linux with a UTF-8 locale it gives Cafe (in the bare C locale, Caf?), but the iconv on macOS and in Windows PHP writes Caf'e, with the accent as a separate character. Python’s popular unicodedata.normalize('NFKD', s).encode('ascii', 'ignore') one-liner removes accents but silently deletes ’, — and € too (Zoes Cafe 5). The JavaScript line strips accents only, leaving quotes and symbols for you to handle. Unidecode also spells out Chinese and Japanese in Latin letters, so 東京 becomes Dong Jing.