Convert String to Hex

Many use charCodeAt() which returns UTF-16 code units, not UTF-8 bytes. So é comes out as E9 (looks fine but isn’t UTF-8) and 😀 comes out as D83D DE00 (UTF-16 surrogate pair garbage). This tool uses TextEncoder for real UTF-8: é → C3 A9, 😀 → F0 9F 98 80. The output matches what other UTF-8 tools (hexdump, xxd, Python bytes) produce.

Yes – click Swap. The reverse parser is lenient: it accepts any of the four output formats, strips 0x prefixes and commas, normalizes whitespace, then decodes as strict UTF-8. 48 65 6C 6C 6F, 0x48,0x65,0x6C,0x6C,0x6F, and 48656c6c6f all give Hello. Odd-digit input (“must have even digit count”) and non-hex characters get specific errors.

The TextDecoder is created with {fatal: true} which throws when bytes don’t form valid UTF-8 – for example a lone continuation byte 80 with no lead byte, or an overlong encoding. Non-strict decoders silently substitute the U+FFFD replacement character ("�") and hide the problem. Strict mode surfaces corruption so you can investigate the source.

Match your destination. Hex dumps and logs: space-separated. CSV imports: comma. Raw byte strings in URLs or DBs: continuous (no separator). C/C++ literal byte arrays: 0x-prefixed like {0x48, 0x65, 0x6C, 0x6C, 0x6F} – paste straight into source code. The output format only affects display; the underlying bytes are identical.

None mathematically – 4F and 4f represent the same byte (79 decimal). The choice is stylistic and depends on your tooling. Unix tools like od and xxd default to lowercase. Many Windows tools and macOS Hex Fiend default to uppercase. C/Rust source code conventionally uses lowercase. Pick whichever your codebase uses.

By iterating codepoints (using for…of on the string) rather than UTF-16 chars. JavaScript stores 😀 internally as the surrogate pair 😀 (two 16-bit code units), but for…of yields it as one codepoint (U+1F600) which TextEncoder then encodes as the correct 4-byte UTF-8 sequence F0 9F 98 80. The grid shows it as one row with byte count 4.

200,000 characters. Beyond that the per-character grid would render too slowly (each row is a DOM node). Hit the cap and you get a clear error pointing to the limit. The grid itself caps at 256 displayed rows with a “… N more characters” note, but conversion still processes the full input – only the grid display is truncated.

No, this tool emits UTF-8 only. UTF-8 is the dominant encoding on the modern web and the natural choice for text-to-hex conversion in 95% of use cases. If you specifically need UTF-16 or windows-1252 hex output (rare), you’ll need a specialized encoder. The reverse direction also only decodes UTF-8 hex.

No. All UTF-8 encoding, hex formatting, decoding, and TSV emission run in your browser using standard TextEncoder/TextDecoder. Open DevTools → Network and confirm zero requests fire – even when you Convert or Download. Safe for sensitive text, credentials being audited, or anything you’d rather not send to a third-party converter.

Yes. Total bundle is about 18 KB. Load once, disconnect, keep using. The conversion is pure local computation via the standard browser APIs – no fonts to fetch, no analytics, no API. Useful for offline debugging, airgapped environments, or when working on data you don’t want to expose.