Character Inspector

What it's for

Look at text one character at a time: each character’s Unicode code point, its UTF-8 bytes and what kind of character it is. It finds invisible and look-alike characters, and explains why two strings that look the same don’t match.

For example, when a password, file name or search won’t match although it looks right, paste it here to find a hidden space, a curly quote, or an accent stored as two characters.

How to use it

Type or paste text. A table lists every character with its code point (such as U+00E9), UTF-8 bytes and type, and the summary counts characters as displayed, code points, UTF-8 bytes and UTF-16 units. It also says whether accents are stored composed (NFC) or decomposed (NFD). Use Convert to composed or Convert to decomposed to switch between the two.

Repair garbled text fixes text whose UTF-8 was read as Windows-1252 or Latin-1, the usual cause of café for café and – for a dash, even when it’s happened twice. Read in another encoding shows what the text’s bytes look like to a program that reads them as Windows-1252, Mac OS Roman, UTF-16 and others, to work out which mistake a garbled file has suffered. International domain names converts a domain such as bücher.example to its Punycode form xn--bcher-kva.example, the form DNS and certificates use, and back.

Example

Type this:

café

The summary reads 4 characters as displayed, 4 code points, 5 UTF-8 bytes, and é is U+00E9, two bytes in UTF-8. Choose Convert to decomposed (NFD) and it becomes 5 code points: an e followed by a combining accent. It looks identical, which is why such strings don’t match.

Good to know

macOS often stores file names decomposed, while most other text is composed. That’s a common reason for file names that look identical failing to match in scripts.