HTTP & Request Debugging
Paste one header block to inspect its fields, parameters and repeated values. Request and response examples show different cases.
Input is processed in this browser; the tool sends no requests and saves no history.
Request / status line
GET /reports HTTP/1.1
example.com
application/json, text/html;q=0.8, */*;q=0.1
application/jsonWeight (q): 1text/htmlWeight (q): 0.8*/*Weight (q): 0.1en, ja;q=0.8, *;q=0
enWeight (q): 1jaWeight (q): 0.8*Weight (q): 0[REDACTED]
first
second
Masks Authorization, Proxy-Authorization, Cookie and Set-Cookie in results and exports. The input stays visible; custom keys and URL secrets are not masked.
One text block, up to 128 KiB UTF-8 and 1000 fields. The first empty line ends the header section.
6 fields · 2 repeated entries · 0 notes
The HTTP headers parser helps when a captured request behaves differently from the request you expected. Keep each occurrence visible before turning a network trace into a test fixture: a repeated field is evidence, not something to discard.
Use HTTP headers to JSON conversion to compare captured values in a script. The report includes ordered entries and grouped arrays, so you can follow the original sequence or retrieve every value for one name.
The HTTP headers parser retains two Set-Cookie lines as two entries, even when their spelling differs. A comma inside Expires stays in its value. Review upstream duplication without a conversion silently choosing the first or last line.
A Content-Disposition parser should distinguish a semicolon inside a quoted filename from a parameter separator. Inspect the raw value beside decoded parameters; the Accept-Language parser likewise exposes weights without rearranging the preferences.
The request example contains Accept, language weights and a fictional bearer token. The response example contains an extended filename and two cookies. Replacing input opens a preview with a character count; Undo restores the previous input and report.
Automatic mode waits 350 ms after edits. Turn it off to inspect a draft, then use Parse headers or Ctrl/Cmd+Enter. Editing input or name display makes the report stale and disables export until the next successful parse.
Masking is on by default. Turning it off reveals the four protected field values in both views and JSON downloads. The Fields and JSON views share one report; Show 50 more only expands the preview, while export always includes all entries.
Each entry records its source line, original spelling and displayed name. The grouped object uses the first displayed spelling as its key and always stores an array, including when the field occurs only once. Names are matched without case distinctions.
Content-Disposition exposes its type, parameter list and suggested filename. A valid filename* takes precedence over filename. UTF-8 and ISO-8859-1 are decoded; ambiguous duplicate parameters or unsupported encodings generate notes instead of an invented filename.
Accept media ranges and Accept-Language ranges retain their input order and parameters. Missing q means 1; q=0 remains visible. An invalid weight becomes null in the detail report and produces a note. The raw field value is kept for diagnosis.
Missing separators, invalid field names, embedded control characters and obsolete folded lines stop parsing with a source location. Malformed parameters inside the three inspected fields produce notes while retaining the field itself. No partial header report is offered after a block syntax error.
No. Content-Type and content-type identify the same field for grouping. Original mode keeps each spelling in the ordered entries; lowercase and title case only change displayed names. Values retain their case.
Ordinary fields remain whole. Only Accept and Accept-Language are split into a separate detail list, with quoted commas respected. Repeated physical lines always remain separate entries, including Set-Cookie.
No. There is no URL fetch action or request replay. Obtain the header block using your own browser or HTTP client, then paste its text here.
A recognizable leading line such as GET / HTTP/1.1 or HTTP/2 200 is saved separately. It is not counted as a field. A block containing just Name: value lines also works.
Yes, if an empty line separates it from the headers. Parsing stops there and reports that subsequent text was ignored. The full input size limit still applies, so remove a large body before pasting.
Found a field that parses incorrectly?
Use the HTTP headers parser on a single captured exchange, review any parameter notes, and export the current JSON report for your next debugging step.