String / Unicode Inspector

How to use it

1. Paste or type text; your last input and any shared permalink are restored, and large pastes are measured off the main thread.
2. Read the four totals on top — grapheme, code point, UTF-16 unit and UTF-8 byte — each copyable on its own.
3. Open any grapheme cluster to see its code points, categories, both lengths and four escapes.
4. Copy one cell or the whole report as JSON, or use the command line past the in-browser limit.

Paste any text and read it one character at a time. This inspector expands every grapheme cluster into its code points and reports four separate lengths — grapheme count, code point count, UTF-16 code units and UTF-8 bytes — beside each character's general category, hex value and escape forms. The whole breakdown is computed in this tab.

How it works

A string can be measured four ways at once. In é😀 a reader sees two things, so the grapheme count is 2; those clusters hold three code points — the letter e, the combining acute accent U+0301, and the face U+1F600. UTF-16 needs four code units because U+1F600 takes a surrogate pair; UTF-8 needs seven bytes. Cluster boundaries follow UAX #29.

That é is why the numbers can surprise you. Written as e followed by U+0301 it is two code points and the NFC line reads no; the precomposed é (U+00E9) is one code point, already in Normalization Form C. This page only reports NFC status — converting between the forms is a dedicated normalization tool's job.

Each code point carries a two-letter general category — Lu for a capital, Mn for the accent, So for a symbol — from the classes in UAX #44. Character names and compatibility folding such as NFKC stay off this page on purpose.

Examples

Hi!

graphemes 3 · code points 3 · UTF-16 units 3 · UTF-8 bytes 3
NFC: yes

grapheme 1 "H" (1 code point)
  U+0048 H  Uppercase Letter (Lu)  UTF-16 1  UTF-8 1  \u{48}
grapheme 2 "i" (1 code point)
  U+0069 i  Lowercase Letter (Ll)  UTF-16 1  UTF-8 1  \u{69}
grapheme 3 "!" (1 code point)
  U+0021 !  Other Punctuation (Po)  UTF-16 1  UTF-8 1  \u{21}

é😀

graphemes 2 · code points 3 · UTF-16 units 4 · UTF-8 bytes 7
NFC: no

grapheme 1 "é" (2 code points)
  U+0065 e  Lowercase Letter (Ll)  UTF-16 1  UTF-8 1  \u{65}
  U+0301 ́  Nonspacing Mark (Mn)  UTF-16 1  UTF-8 2  \u{301}
grapheme 2 "😀" (1 code point)
  U+1F600 😀  Other Symbol (So)  UTF-16 2  UTF-8 4  \u{1F600}

👍🏽

graphemes 1 · code points 2 · UTF-16 units 4 · UTF-8 bytes 8
NFC: yes

grapheme 1 "👍🏽" (2 code points)
  U+1F44D 👍  Other Symbol (So)  UTF-16 2  UTF-8 4  \u{1F44D}
  U+1F3FD 🏽  Modifier Symbol (Sk)  UTF-16 2  UTF-8 4  \u{1F3FD}

Every example on this page is executed against the same engine that powers the tool, on every build.

Common errors

Input is larger than the 1 MB limit for this tool. For bigger inputs use the CLI: uni identify < input.txt

The whole input is declined, never partly measured — a partial read would misreport every total. arp242/uni inspects any size locally, and nothing is uploaded.

This input contains an unpaired surrogate code unit. It is shown as the replacement character U+FFFD (�) and counts as 3 UTF-8 bytes.

A lone surrogate usually comes from a string sliced through the middle of an emoji. The note says it is present and how it was counted, without hiding or rejecting it.

Showing the first 20000 of 20001 code points in the table below — the four totals above still cover the entire input.

Only the table is capped, so a long paste never freezes the page; the four totals still cover everything, and the JSON report has the rest.

Keyboard shortcuts

⌘/Ctrl + Enter

Measure the current text again immediately.

⌘/Ctrl + S

Save the breakdown as a .json file, not the web page.

⌘/Ctrl + C

With nothing selected, copy the JSON breakdown; a selection copies itself.

⌘/Ctrl + Shift + C

Copy the JSON breakdown whether or not anything is selected.

t

Step to the next theme (idle while a field has focus).

? or ⌘/Ctrl + /

Open the shortcut help panel.

Esc

Dismiss the top overlay and return focus to its opener.