Method and bounds

Inspect one form-query model.

Queryfold follows the WHATWG application/x-www-form-urlencoded parsing model within a deliberately narrow input domain. It accepts printable ASCII query text. Unicode content can still be inspected when represented by UTF-8 percent bytes, such as caf%C3%A9. This is a form-query inspector, not a generic decoder for URL paths or every server encoding.

From raw segments to ordered pairs

The page removes at most one initial question mark as an explicit convenience. It then separates nonempty ampersand-delimited segments. Within each segment, only the first equals sign divides the raw name and value. A missing equals sign gives an empty value, while the separate “Equals present” flag stays false. Empty segments between ampersands are ignored; empty names are not.

Next, plus becomes space. A percent sign consumes two following characters only when both are hexadecimal digits. Those bytes are decoded as UTF-8. The Encoding Standard’s decoding hook uses replacement characters for invalid sequences and keeps a leading BOM as data. Consequently, decoding is not always lossless. The raw token columns remain available so that distinction is visible.

Keep duplicates without guessing server behaviour

The result is an ordered list, not an object indexed by names. Repeated decoded names remain separate, even when their raw spelling differs. For example, %61=1&a=2 contains two pairs whose decoded name is the same lowercase a. Name equality is exact and case-sensitive. The page does not sort, normalize, deduplicate or convert values to numbers, dates or booleans.

The limits are part of the interface

Input is limited to 100,000 printable ASCII characters and 2,000 nonempty pairs. Raw tabs, line breaks and non-ASCII characters are rejected. A complete JSON result above 1,000,000 UTF-16 units is rejected before exports become available. The preview shows 100 rows per page; saved results contain every current row.

Paste only the query portion. A literal # and any second ? are data; the page does not extract a URL’s fragment or destination. It does not visit a link, validate credentials, remove tracking, judge safety or predict a remote application’s duplicate-key policy. Textarea handling may normalize pasted line endings, so this tool promises interpretation of the received string, not preservation of original clipboard or file bytes.