Unicode Code Point & UTF-8/UTF-16 Inspector

Understanding the Unicode code point inspector — concept and principle

Every character is stored as a number called a Unicode code point, written U+ followed by at least four hex digits. An encoding then turns that number into bytes, usually UTF-8 (files, web, APIs) or UTF-16 (JavaScript, Java, Windows). This inspector shows, per character, the code point, the UTF-8 bytes and the UTF-16 units, which helps when debugging garbled text or checking what a database will store.

The emoji lies outside the BMP, so it needs 4 UTF-8 bytes and a UTF-16 surrogate pair.

How it is calculated

Code point: the number assigned by the Unicode Standard, U+0000 to U+10FFFF.
UTF-8: 1 byte up to U+007F, 2 bytes up to U+07FF, 3 bytes up to U+FFFF, 4 bytes above. The first byte signals the length (0xxxxxxx, 110xxxxx, 1110xxxx, 11110xxx); each following byte is 10xxxxxx and carries 6 bits.
UTF-16: one 16-bit unit inside the Basic Multilingual Plane (BMP, up to U+FFFF). Above it, subtract 0x10000 and split the 20 bits into a high surrogate 0xD800 + top 10 bits and a low surrogate 0xDC00 + bottom 10 bits.
Counts: code points / UTF-16 units (JavaScript .length) / UTF-8 bytes.

The principle

Unicode separates identity from storage: the code point says which character it is, the encoding decides how many bytes to spend. UTF-8 keeps ASCII as one byte, and because continuation bytes always begin with 10, a reader can resynchronise anywhere in a stream. UTF-16 reaches the full 1,114,112 code points by using the reserved range U+D800–U+DFFF as surrogate pairs.

Worked example

Input: 한😀A, three characters.
한 = U+D55C. Its bits 1101 010101 011100 fill the 3-byte pattern as ED 95 9C; in UTF-16 it is one unit, D55C.
😀 = U+1F600, outside the BMP. UTF-8: F0 9F 98 80. UTF-16: 0x1F600 − 0x10000 = 0xF600, top 10 bits 0x3D, bottom 10 bits 0x200, giving D83D DE00.
A = U+0041: UTF-8 41, UTF-16 0041.
Totals: 3 code points, 1 + 2 + 1 = 4 UTF-16 units, 3 + 4 + 1 = 8 bytes, shown as 3 / 4 / 8.

How to read the result

A UTF-16 length larger than the character count means the text contains characters beyond the BMP, usually emoji. Korean syllables (U+AC00–U+D7A3) always take 3 bytes in UTF-8 and one UTF-16 unit.

Common mistakes

Assuming one character is one byte, or that JavaScript .length counts characters.
Cutting a string inside a surrogate pair or a multi-byte UTF-8 sequence.
Treating a skin-tone or family emoji as one code point; these are sequences joined by modifiers or U+200D.

Keep in mind: "Characters" here means code points. What a reader sees as one character (a flag, a decomposed Hangul syllable) can be several code points, so use a grapheme-aware library when limits are about visible characters.

Frequently asked questions

Why does MySQL reject emoji in a utf8 column?

Legacy utf8 (utf8mb3) stores at most 3 bytes per character, so 4-byte characters do not fit. Use utf8mb4.

Is UTF-16 always 2 bytes per character?

No. Characters above U+FFFF take 4 bytes as a surrogate pair.

Which encoding should I use for new files?

UTF-8 is the default for the web and most tools; it is compact for ASCII and has no byte-order issue.

Sources: The Unicode Consortium, The Unicode Standard, Chapters 2 and 3 · IETF RFC 3629, UTF-8, a transformation format of ISO 10646 · IETF RFC 2781, UTF-16, an encoding of ISO 10646

Checked 8 Nov 2026. Standard formulas and RFC/IEEE definitions, computed locally in your browser — nothing you type is sent anywhere.

Your numbers