Character and Encoding Inspector

How to use

1. Paste the text you want to inspect.
2. Choose whether the table shows only unusual characters or every character.
3. Read the findings, then the table of code points with their UTF-8 bytes and names.
4. Copy the text tagged, escaped or cleaned, or try the mojibake repair if the text looks garbled.

What it checks and how it works

The tool lists each code point in the text with its U+ number, its UTF-8 bytes, its Unicode general category, read from the JavaScript engine's own Unicode tables, and, for invisible, format, space and bidirectional control characters, a name and a kind. Names are bundled only for those characters. For any other character the table gives the general category and not a Unicode character name. Bidirectional classes and controls follow Unicode Standard Annex #9, using the bidi-js package, and unbalanced controls are reported. Unterminated bidirectional controls in source code are the basis of the Trojan Source attack described at trojansource.codes (CVE-2021-42574).

The text can be shown with unusual characters tagged, written with Unicode escapes, or cleaned. Cleaning can keep the zero-width non-joiner and joiner, which Arabic and Persian writing use. The tool can also try to repair mojibake, text decoded in the wrong encoding, by reading it as windows-1252 bytes and decoding those bytes as UTF-8, as the WHATWG Encoding Standard defines.

Limits: a repair is a guess to check by eye, and the tool does not normalize or spell-check text. When you ask to see every character, a maximum of 300 are listed.