How to read the result
The label breakdown shows each part of the name between the dots, with its encoded length. One part can't be longer than 63 bytes.
For a name with non-ASCII characters, the character table lists each one with its code point, Unicode name, and writing system. A line under the table says whether any part mixes writing systems, which is the usual sign of a lookalike.
аpple.com, with a Cyrillic first letter
Example
The ASCII form is xn--pple-43d.com. The character table shows the first letter as U+0430 CYRILLIC SMALL LETTER A, and the result flags that the label mixes Cyrillic and Latin.
What is punycode?
Punycode is the way a domain name with non-ASCII letters is written in plain ASCII, so DNS can handle it. Each encoded part starts with xn--. münchen.de and xn--mnchen-3ya.de are the same domain, written two ways.
RFC 3492 defines the encoding, and RFC 5890 sets the rules for which names are valid. Not every Unicode string can be a domain name, so the result lists any rule a name breaks.
How do you spot a punycode phishing domain?
Look for letters from another writing system that copy familiar ones. Latin a is U+0061, and Cyrillic а is U+0430. They look the same, but they make a different domain. This trick is called a homograph attack.
Most lookalikes mix writing systems inside one part of the name, so the tool flags that and shows each character. A mixed name is a reason to look closer, not proof of fraud. A name written entirely in Cyrillic or Greek is normal where those languages are used.
What won't this catch?
Lookalikes made only of ASCII letters. paypaI.com with a capital i, or rn standing in for m, use ordinary Latin letters, so no writing system check can flag them. No warning doesn't mean the name is safe.
Your browser also decides when to show the readable form and when to show xn--. The rules differ between browsers and change over time. Chromium's IDN display policy is one example.
Why is my converted domain name too long?
Encoding makes non-ASCII names longer. 例え.jp has 2 letters before the dot, but its encoded form xn--r8jz45g has 11. Each part has a 63-byte limit after encoding, and the whole name has a 253-byte limit. The label breakdown shows each part's encoded length.
Does punycode apply to the whole URL?
No, only the domain name. The path and query use percent encoding, such as %20 for a space. When you paste a full address, only its hostname changes.
If a link looks suspicious, the domain age checker shows when the name was registered.