Hex to Text

What this tool does

Paste hexadecimal in whatever shape you found it and this turns it back into readable text. It does not insist on a format: spaces, commas, colons, dashes, newlines, 0x prefixes, \x escapes and one continuous run are all accepted, and the tool tells you which shape it recognised.

The bytes are then decoded as UTF-8 by default. When that fails — because the data is Latin-1, UTF-16 or not text at all — it says so rather than producing a screen of question marks, and lets you switch interpretation or download the raw bytes.

Common uses

• Reading a packet capture or a serial log, where the payload is a run of bytes and only some of it is text.
• Turning a Postgres bytea or MySQL BINARY value back into the string that was stored.
• Checking what an embedded device actually transmitted, byte by byte, against what the protocol document says.
• Decoding a value from a hex-encoded config file or a firmware dump.
• Confirming whether a file starts with a known magic number — 89 50 4e 47 is a PNG, 25 50 44 46 is a PDF.

A short example

These four inputs are identical as far as this tool is concerned:

48 65 6c 6c 6f
48656c6c6f
0x48, 0x65, 0x6C, 0x6C, 0x6F
48:65:6c:6c:6f

All four decode to:

Hello

When decoding goes wrong

Two failures account for most of them, and each has a specific fix:

• An odd number of digits. Every byte is exactly two digits, so an odd count means one is missing — usually a dropped leading zero, where 0f was written as f.
• Bytes that are not valid UTF-8. A byte in the 0x80–0xFF range only makes sense as part of a multi-byte sequence. If those sequences are malformed, the data is probably Latin-1 (where every byte is its own character) or genuinely binary.

A strong hint sits in the stats above: a high proportion of printable ASCII means text, while alternating zero bytes mean UTF-16 — 48 00 69 00 is "Hi" in little-endian UTF-16, and decoding it as UTF-8 produces "H\0i\0".

Worth knowing

Hex is a view, not a format — it says nothing about what the bytes mean. The same c3 a9 is "é" in UTF-8, "é" in Latin-1, and a fragment of a compressed stream in a gzip file. Choosing the interpretation is your job, and the tool's analysis is a hint rather than an answer.

One convenience worth knowing about: a leading # is stripped, so a CSS colour like #ff8800 parses as the three bytes 255, 136, 0 — handy when checking a value against an RGB triple.