Unicode and Internationalization
Inspect text, decode hexadecimal values and compare the units that make up a character.
Preview shows at most 4,000 output characters; copying and downloading include the full output. Text preview uses escaped diagnostic markers.
Input stays in this browser and is not uploaded.
Text: up to 10,000 UTF-16 units. Encoded input: up to 200,000 characters, decoding to at most 10,000 units. Code points use U+HEX, 0xHEX or HEX; bytes use two HEX digits, optionally 0x. Separate tokens with whitespace or commas. No escapes or HTML entities.
Names: Unicode 17.0.0. Graphemes: browser Intl.Segmenter; its Unicode version is not exposed and may differ. Names and control aliases remain in official English.
Byte order is explicit. Initial U+FEFF is preserved; no BOM is detected, removed or added. Input is not normalized. Untouched CR/CRLF received from the browser are preserved; the textarea displays line endings as LF.
Detail offsets are zero-based UTF-16 units with exclusive ends. Error line/column and byte positions start at 1; missing bytes point just past the input. Scalar counts exclude isolated surrogate units.
Enter text or load an example.
A Unicode inspector helps when the length of a string disagrees with what you see. An accented letter can contain a base letter and a combining mark; a family emoji can contain several people and joiners. Counting the visible shape alone misses that structure.
This Unicode character inspector keeps the original sequence intact while showing names and encoding details. Use code point lookup for a pasted symbol, or start from hexadecimal bytes when checking a value copied from a log or protocol trace.
The Unicode inspector separates scalar code points, UTF-16 units and graphemes. Compare a precomposed accent with a letter followed by a combining mark: their appearance can match while their stored values and unit counts differ.
Select a byte input format and inspect the decoded character, then export it in another encoding. Explicit UTF-16 and UTF-32 byte order lets you compare the two orders without guessing from a file header. Invalid input stops at a reported position.
Text is the default. Code points accept U+0041, 0x41 or 41; a byte is exactly two hexadecimal digits, optionally prefixed by 0x. Separate tokens with whitespace or commas. A JavaScript escape is not a code-point token.
Automatic inspection waits 250 ms after an edit. Turn it off to use Inspect or Ctrl/Cmd+Enter. A changed input makes the old report stale and disables export until inspection finishes. Example and Clear preview replacements when text already exists; Apply commits and Undo restores.
Switch between code points and grapheme boundaries; reveal another 50 entries as needed. The output preview stops at 4,000 characters, but copying or downloading exports everything. Diagnostic markers in text preview are display aids; valid text export contains the actual characters.
Names come from Unicode 17.0.0 data, including algorithmic naming ranges. Official English names stay unchanged across languages. Control aliases have their own label; private-use, noncharacter and unassigned values are identified without inventing names.
Read UTF-8 bytes, UTF-16 code units and UTF-32 scalar values beside each entry. Output options include explicit little-endian and big-endian byte sequences. The JSON report preserves the inspection structure for a bug report or a follow-up script.
Detail ranges count UTF-16 units from zero and exclude the end offset. Error lines, columns and byte positions start at one. These are storage coordinates, so they need not match a cursor position measured in visible glyphs.
Surrogate code units occupy the D800–DFFF range. UTF-16 pairs a high unit with a low unit to represent a supplementary scalar value. A lone unit has no valid scalar interpretation; it is shown here for diagnosis and is never silently encoded as a replacement character.
UTF-16 represents a scalar value using one or two 16-bit code units. A supplementary emoji uses a pair, so one code point can contribute two units and four bytes. The units view and the byte output are deliberately labeled differently.
No. Unicode defines a code space through U+10FFFF. UTF-8, UTF-16 and UTF-32 are ways to represent scalar values from that space. The number in an encoding name describes its code-unit width, not a limit on the number of characters Unicode can represent.
One visible unit can combine several values: a letter and accent, an emoji and modifier, or a sequence joined by U+200D. The grapheme view groups these using the browser’s segmenter, while the code-point view keeps each constituent visible.
Yes. Paste a string within the stated limit. Each code point appears in the detail list, and the report includes the complete result even when only the first page is visible. Start with a short fragment if you need to isolate one unexpected value.
Does a name or boundary look unexpected?
Use the Unicode inspector to compare the smallest example that reproduces your encoding problem, then copy the representation your next step needs.