Worked examples

Small details, kept visible.

These examples describe the declared form-query interpretation. Paste them into the calculator and choose Inspect query. They are query portions rather than full URLs. No example requires visiting a destination or granting access to another account.

Repeated names and empty values

?tag=blue&tag=green&empty=&flag&name=Ada+Lovelace produces five ordered pairs and four distinct decoded names. Only one name, tag, is used more than once. Its values blue and green remain separate and in order. Both empty and flag have empty values, but “Equals present” is true for empty and false for flag. The last value decodes to Ada Lovelace.

A plus is not always a space

In x=%2B+%25, the decoded value is a literal plus, a space, and a percent sign: "+ %". The raw token is retained alongside it. This is why simply replacing every visible plus after decoding would be the wrong operation: the order of the transformations matters.

The first equals sign is structural

a=b=c has one pair. Its name is a and its value is b=c. Meanwhile, &&a=1&& also contains one pair because empty ampersand segments are ignored. =value&=second contains two pairs with the same empty decoded name. Empty strings are shown with quotes so they do not vanish visually.

Malformed escapes and incomplete UTF-8

bad=%G1&short=%&utf8=%E2%82 keeps the malformed percent spellings in the first two values. The incomplete UTF-8 sequence in the third is decoded with a replacement character. This is not proof of what the original intended text was. Use the raw token to see which bytes were entered.

Boundaries that are intentional

bom=%EF%BB%BFx retains a leading U+FEFF in the decoded value; the preview escapes it so it is visible. ??a=1 removes only the first question mark, leaving the name ?a. A literal # remains data. To inspect Unicode text, use percent bytes such as unicode=caf%C3%A9+%E9%9B%AA, which decodes to café followed by a space and 雪.