See every code point in a string with its category, script and escapes, all four of its lengths, and what each normalisation form would change.
Waiting for text
Every length a string has, every code point in it, and what normalisation would change.
The report appears here
Every length the string has, each code point with its category and script, and what normalisation would change.
Unicode Inspector runs entirely on your device — nothing is uploaded.
You’re on the Free plan. Compare plans
How long is this string has four correct answers, and that is the reason this tool exists. A three-person family emoji is eight UTF-16 code units, which is what JavaScript's length property returns. It is five code points, which is what a Unicode-aware loop iterates. It is eighteen UTF-8 bytes, which is what a column sized in bytes counts. And it is one character, which is what the person reading it sees. A tool that reported a single length would be answering a question you did not ask.
Every code point is listed with its category, its script, its UTF-8 bytes, its UTF-16 code units, an escape sequence and an HTML entity. Composite characters are grouped as one row with their parts beneath, so it is visible when a single visible character is in fact a base letter plus two combining marks plus a variation selector.
The character count comes from the browser's own grapheme segmenter. Where that is unavailable the engine falls back to counting code points and the page says it has, rather than presenting an approximate count as an exact one. That matters because the fallback over-counts exactly the cases people come here to understand.
Normalisation is the other half. Two strings that look identical can fail an equality check because one holds a precomposed character and the other a base letter with a combining accent. All four forms are shown with their lengths, which form the input already matches, and whether the compatibility forms differ — because those are lossy by design, folding a ligature into its letters and a circled digit into a plain one. Right for a search index, wrong for anything you store and show back to someone.
On the category names: Unicode defines thirty two-letter general categories, and this groups them into eleven readable ones, so letter covers uppercase, lowercase, titlecase, modifier and other letters without distinguishing them. That is an approximation chosen for readability, and it is stated where the categories are shown rather than left to be inferred.
Because each measures something different. JavaScript's length counts UTF-16 code units, a for-of loop counts code points, a byte-sized database column counts UTF-8 bytes, and a reader counts the characters they can see. For plain ASCII all four agree; for anything else they do not.
The technical name for one character as a reader perceives it, which may be several code points: a letter plus a combining accent, a flag made of two regional indicators, or an emoji family joined with zero-width joiners. The character count here counts these.
Because that browser has no Intl.Segmenter, so the count falls back to code points. The fallback over-counts exactly the composite characters you would come here to look at, so the page labels the number rather than presenting it as exact.
It rewrites a string into a canonical form so two strings that look the same compare equal. NFC composes characters, NFD decomposes them. NFKC and NFKD also fold compatibility characters, which is lossy — a ligature becomes its letters and a circled digit becomes a plain one.
NFC for anything you store and display, since it is the web's default and preserves the text. NFKC for a search index or a comparison key, where folding a ligature into its letters is what you want. Not NFKC for data you will show back to someone, because the change is permanent.
They are readable groupings of them. Unicode defines thirty two-letter categories; these are eleven names covering those, so Letter spans several official categories without separating them. The page says so beside the list.