CJK text guide · October 28, 2026
Does Claude Text Cleaner Remove Ideographic Variation Selectors?
Yes. The browser cleaner removes U+E0100 through U+E01EF from pasted text and counts them as “Variation selectors.” In a live check, 龍 + U+E0100 became 龍: two Unicode code points became one, and the report showed one removal. The base ideograph stays, but a requested glyph variant may be lost. Keep the original if its exact form matters.
龍 + U+E0100Base ideograph followed by a variation selector.龍Variation selectors: 1. The base character remains.Why can an invisible selector matter?
A Unicode variation sequence is a base character followed by a selector that requests a particular appearance. Registered ideographic variation sequences use selectors in the U+E0100–U+E01EF range. Support varies by font and software, so two versions can look alike in one editor and differ in another. The cleaner applies a code-point rule; it does not determine whether a pair is registered or whether a recipient can display the requested form.
A public English-language question about Chinese and Japanese character appearance illustrates why writers may care about a particular glyph in mixed-language text. That discussion is about text appearance generally, not a report of this site's behavior.
How to check and preserve your text
- Save the original. Work on a copy if names, quoted text, or typography require an exact glyph.
- Compare code points. Paste the text into the cleaner. A “Variation selectors” count above zero includes both this supplementary range and the U+FE00–U+FE0F range used for other variation sequences.
- Review in the destination. Compare original and cleaned text using the intended font and application. If the form matters, keep the original sequence or restore it manually.
In the live check, pasting 龍󠄀 produced 龍; original length 2, clean length 1, removed characters 1, changed categories 1, and “Variation selectors” 1. These lengths count Unicode code points, not bytes or UTF-16 code units. The page only cleans text pasted in the browser. A removal count does not establish where the text came from, identify a watermark, or predict an AI detector result. For a different use of variation selectors, see the emoji guide.