Convert text to its hexadecimal bytes or read hex back as text. The encoding is UTF-8 and it is named, because the right answer depends on which one you mean.
Waiting for text
Text is encoded as UTF-8 bytes, then written in hexadecimal. The answer depends on the encoding, so this one is named.
The result appears here
Every byte of the UTF-8 encoding, written in hexadecimal.
Hex to Text runs entirely on your device — nothing is uploaded.
You’re on the Free plan. Compare plans
The hex of this text is not a well-defined question until you say which encoding. The letter e-acute is C3 A9 in UTF-8, E9 in Latin-1, and either 00E9 or E900 in UTF-16 depending on byte order. This converts UTF-8 and says so everywhere, because a converter that quietly picks an encoding gives an answer that is right for one caller and wrong for the next.
The byte count is therefore not the character count, and the tool shows both. Anything outside ASCII takes more than one byte — two for most accented letters and Greek or Cyrillic, three for most of the world's other scripts, four for emoji and the rest of the supplementary planes. That difference is the usual reason a value that looks short enough for a database column is not.
Decoding is strict. Bytes that are not valid UTF-8 are rejected with an explanation rather than decoded with replacement characters, because a lossy decode looks exactly like a successful one until something later breaks. An odd number of hex digits is rejected too, with the count, since a byte is always two digits and an odd total means a digit went missing in copying.
On the way in, separators are forgiven: spaces, colons, commas, hyphens, underscores and 0x prefixes are all stripped before the digits are read, so a hex dump pasted from a packet trace, a certificate fingerprint and one unbroken run all work. Anything that is not a separator and not a hex digit is an error rather than something to skip, because skipping it would change the bytes.
On the way out there are five groupings, each for a different place the result is going: spaces for reading, nothing for a single token, colons for a fingerprint, 0x prefixes for pasting into source, and sixteen bytes per line for something that looks like a hex dump. Nothing is uploaded.
UTF-8, both ways, and it is stated on the page. The hex of a string depends entirely on the encoding, so a tool that does not name one cannot be relied on for anything outside ASCII.
Because UTF-8 uses more than one byte for anything outside ASCII — two for most accented letters, three for most other scripts, four for emoji. That gap is usually why a value overflows a column sized in bytes.
Either it contains a character that is neither a separator nor a hex digit, or the digit count is odd. A byte is exactly two digits, so an odd total means one was lost. Both errors name the cause.
Those bytes are not a legal UTF-8 sequence. They may be text in another encoding, or not text at all. The tool declines rather than substituting replacement characters, because a lossy decode is indistinguishable from a correct one at a glance.
Spaces, colons, commas, hyphens, underscores and 0x prefixes are all ignored, so a certificate fingerprint, a packet dump and one long run of digits all work without cleaning up first.
No. Both represent bytes as text, but hex uses two characters per byte and is readable one byte at a time, while base64 uses roughly four characters per three bytes and is more compact. Hex is for inspecting bytes; base64 is for transporting them.