Punycode Converter

This converter implements the Punycode Bootstring algorithm (RFC 3492) directly in the browser rather than calling a library, using the standard parameter set hard-coded in the component: base 36, tmin 1, tmax 26, skew 38, damp 700, an initial bias of 72, an initial code point of 128, and a hyphen as the delimiter. The adaptive bias function and the main encode/decode loops follow the spec, including overflow guards that throw when an intermediate value would exceed Number.MAX_SAFE_INTEGER. All work happens client-side as you type, so domains never leave your machine, and a mode toggle flips between Unicode-to-Punycode and Punycode-to-Unicode with the input/output panels swapping which side is editable.

Crucially, encoding is applied per label, not to the whole string. The domain is split on dots, and each label is handled independently: a label is only Punycode-encoded if it contains a character outside the ASCII range (the regex /[^\x00-\x7F]/), in which case it is lowercased and prefixed with the xn-- ACE marker; pure-ASCII labels such as the .jp or .com suffix pass through untouched. Decoding is the mirror image - only labels that begin with xn-- (case-insensitively) are run through the Bootstring decoder, and any label whose Punycode is malformed is silently returned as-is rather than erroring the whole domain, because the decode is wrapped in a per-label try/catch. So 日本語.jp becomes xn--wgv71a119e.jp and münchen.de becomes xn--mnchen-3ya.de, while the TLD label is left as plain ASCII.

Because the tool surfaces the literal xn-- form of any Unicode domain, it is useful for inspecting suspicious internationalized links: a label that visually reads like a familiar brand but contains a Cyrillic or other look-alike character will encode to a distinct xn-- string, which is the same canonical form browsers display for confusable labels. Note that this is a faithful round-trip codec only - it performs Punycode encode/decode and the xn-- prefix logic, but it does not run IDNA2008 or UTS-46 mapping, NFC normalization, or the script-mixing and confusable-detection rules that real registrars and browsers enforce, so a clean conversion here is not a validity or safety verdict. Six preset example domains spanning Japanese (日本語.jp), German (münchen.de), Russian (москва.рф), Chinese (中文.com), Arabic (القاهرة.مصر), and Korean (한국.한국) scripts are provided to load and convert with one click.