Hexspeak Converter

Two precise ways to make hexspeak

Hexadecimal is base 16, using 0–9 and A–F. For hexadecimal digits d₀ … dₙ, the represented integer is:

N = d₀ × 16ⁿ + d₁ × 16ⁿ⁻¹ + … + dₙ

In Decimal Integer mode, the converter accepts a non-negative decimal integer up to 256 digits and produces its raw uppercase hexadecimal value without relying on PHP’s fixed-width integer range. It then tests the classic readable convention used by this tool: 0 → O, 1 → I, while A–F remain letters. Any raw digit from 2 through 9 makes that classic reading invalid. For example, decimal 12648430 becomes raw C0FFEE, which reads as COFFEE; decimal 4919 becomes 1337, which cannot be rendered under these strict classic rules.

Latin Phrase mode asks a different question: can the supplied phrase be represented with hexadecimal glyphs? Exact mode accepts only A–F and literal digits. Conservative mode additionally maps O → 0, I or L → 1, S → 5, T → 7, and G → 6. These substitutions are deliberately narrow. Unsupported characters are listed, never discarded, and output formatting is available only when the whole phrase is valid.

Tokens such as DEADBEEF, CAFEBABE, BADF00D and C0FFEE are memorable because they resemble words while remaining valid hexadecimal. One documented production example is 0xCAFEBABE: the Java Virtual Machine Specification defines it as the magic number at the start of a Java class file. That does not make every clever-looking value a standard or safe identifier; check the requirements and width of the system where you plan to use it.

Hexspeak is not text-to-HEX

A text-to-HEX encoder converts encoded bytes: UTF-8 Coffee, for example, becomes 43 6F 66 66 65 65. Given the character encoding, that byte representation can be decoded. Hexspeak instead creates a visual word from a restricted character set. Substitutions such as O → 0 and both I/L → 1 are ambiguous, so phrase stylization is not reversible. It is also narrower than leetspeak, which may use many characters that are not valid hexadecimal digits.

Input is limited to 256 characters, and a phrase may contain at most 32 words. Phrase mode supports unaccented ASCII Latin letters, digits and spaces only; transliterate other scripts and remove diacritics before testing them. Decimal and phrase values are processed by the server and may also appear in multi-step funnel URLs, so do not enter secrets, credentials or private identifiers.

Sources: NIST on hexadecimal notation and four-bit digits · Java Virtual Machine Specification, class-file magic