Convert UTF-8 to Arbitrary Base

Each individual UTF-8 byte (value 0-255) gets written in your chosen number base. Base 16 produces hex bytes (00-FF). Base 2 produces 8-bit binary. Base 36 packs each byte into 2 chars using digits 0-9 and letters A-Z. The tool is NOT changing the underlying bytes – only how they’re written.

JavaScript’s Number.prototype.toString(radix) supports 2-36 because that’s the alphabet of 0-9 (10 digits) + A-Z (26 letters). Above 36 you’d need to define your own alphabet (Base58, Base64, Base85 all use distinct character sets – those are separate tools).

A byte’s max value (255) needs different widths in different bases: 8 digits in binary (255 = 11111111), 3 in octal (377), 2 in hex (FF), 2 in base 36 (73). Padding fills shorter byte representations with leading zeros so reverse decoding can split fixed-width tokens unambiguously. With padding off, byte 5 in base 2 is just 101 (ambiguous with byte 0x101 = 257 which can’t exist) and reverse needs explicit separators.

Encoded byte-by-byte exactly as UTF-8 produces them. 🌍 (U+1F30D) is 4 UTF-8 bytes F0 9F 8C 8D in hex; in base 2 that’s four 8-bit tokens; in base 36 it’s 6Y 4F 4D 4D. The per-character grid shows the multi-byte expansion.

Yes – strictly. The tool runs the parsed bytes through TextDecoder('utf-8', {fatal: true}), so invalid sequences (truncated multi-byte chars, overlong encodings, bytes encoding surrogate codepoints) throw an explicit error rather than substituting U+FFFD replacement characters. If decoding fails, the message usually means you have the wrong base or corrupted input.

Each token must be a single byte (0-255). If your token parses to a larger value in the chosen base – say FFF in hex = 4095, well above 255 – that’s a clue the input was meant for a different base, or wasn’t padded properly, so neighbouring tokens got merged.

Mostly education (showing how base conversion generalises) or constraint games (encoding data in a system that only allows certain characters). Base 36 specifically packs bytes more compactly than hex – 2 chars per byte at most, same as hex, but with 26 extra symbols in the alphabet.

No – Base64 isn’t a mathematical radix; it’s a specific encoding spec (3 input bytes → 4 output chars from a 64-char alphabet including + / or URL-safe - _). For Base64, use the dedicated converter in this suite.

No. TextEncoder and TextDecoder run entirely in the browser. About 18 KB of code.

200,000 characters. Tokens grow with smaller bases (base 2 produces 8× the chars), so this cap protects against tab-freezes on huge inputs.