A visible character can contain several code points
Two strings can look identical while containing different sequences. The letter é may be stored as one precomposed character or as e followed by a combining acute accent. A person sees one letter in both cases. Software performing a raw string comparison can see a difference. This inspector shows code points and UTF-16 offsets so you can locate that difference instead of repeatedly retyping a suspicious word.
Grapheme clusters describe the groups people usually perceive as characters. They are useful when inspecting emoji sequences and combining marks. A count of UTF-8 bytes answers a different question: how much space the text occupies when encoded. The table lists the byte representation of each code point. Browser segmentation support determines whether a grapheme total can be shown.
Not every invisible character is a mistake. A nonbreaking space prevents a line break at a particular place. Joiners can affect the shaping of Indic text or combine emoji into a single presentation. Directional controls influence how mixed-direction text is displayed. The inspector labels common controls and classifies other format characters, but it does not claim to provide every official Unicode character name or detect malicious intent.
Normalization is an explicit preview. NFC and NFD handle canonical composition and decomposition. NFKC and NFKD also apply compatibility mappings, which may remove distinctions you intended to preserve, such as full-width forms. Review the result before using it. The original remains available, and no blanket invisible-character removal is performed. Hidden characters are not reliable evidence that a text was written by AI.