Convert UTF-8 to ASCII

They are combined before any ASCII folding happens. A sequence like xC3xA9 (the two bytes for é) goes through TextDecoder('utf-8'), which merges multi-byte sequences into single Unicode codepoints – so é is treated as one character, not the mojibake pair é.

Unicode NFD normalization decomposes accented characters into base + combining marks (é = e + U+0301 acute), then we drop everything in the combining-mark range (U+0300-U+036F). Result: café → cafe, naïve → naive, résumé → resume. Doesn’t transliterate non-Latin scripts (Cyrillic, CJK, Arabic) – those still need ? via the replace fallback.

Six: xHH C-style escapes, uHHHH JS/Python escapes, %HH URL-encoded, raw space-separated hex, raw space-separated decimal, and HTML numeric entities (H decimal or H hex). Detection runs in that priority order: the first regex that matches wins. So if your input mixes formats, the first detected wins and the rest gets parsed by that format’s rules – which may fail.

The decoder runs in fatal mode and throws – for example a lone xC3 (high byte expecting a continuation) errors with “Invalid UTF-8 byte sequence”. The old version would silently produce Ã (Latin-1 byte 0xC3) which masks the underlying corruption.

Because u escapes are by definition Unicode codepoints (16-bit ranges per JS/Python convention), not bytes. The tool converts each uHHHH to the corresponding Unicode character, then encodes the resulting text as UTF-8 internally before ASCII-folding. é → é → 2 UTF-8 bytes → fold per chosen mode.

Both &#NN; (decimal codepoint) and &#xHH; (hex codepoint) are extracted via regex and translated codepoint-by-codepoint. Non-numeric named entities like © are NOT supported here – use the dedicated HTML entity decoder for those.

Both remove non-ASCII content, but replace preserves character positions (café → caf?, 4 chars) while strip shortens output (café → caf, 3 chars). Replace is safer when downstream code depends on character positions.

No. Everything runs in the browser. About 22 KB.

200,000 characters. Long inputs throw an explicit error instead of freezing the tab.

Yes – click Swap. Choose any of 6 output formats including xHH, uHHHH, URL-encoded, raw hex, raw decimal, and HTML-entity (which leaves ASCII alone and only entity-encodes non-ASCII for safe HTML embedding).