How to use HTTP Content Negotiation Tester
- Enter Accept, Accept-Language and Accept-Encoding preferences. Leave a field blank to model an absent header, not a captured empty-header distinction.
- Use editable representation rows or the text list to supply media type, language, coding and an optional label. A dash represents no language or identity coding.
- Choose listed-order tie breaking or an unresolved tie. Inspect the updated quality scores and exclusion explanations.
- Download the ranking/report if useful, then compare with the actual server. This page does not fetch or transform a representation.
Example: HTTP Content Negotiation Tester
Rank the supplied English gzip HTML first under the example’s media, language and coding preferences.
Options
- Quality and specificity
- A more specific matching media range determines that representation’s quality under the selected rules. q=0 excludes an applicable choice.
- Tie policy
- Declared order can resolve equal model scores, or the model can leave a tie unresolved. A real server can apply its own policy.
- Accept
- One or more Accept field values, one per line. Media ranges may carry parameters and a q value, for example "text/html;level=1;q=0.8". Up to 24 ranges.
- Accept-Language
- One or more language ranges such as "en-GB" or "*", each with an optional q value. The local syntax subset accepts a 1–8-letter first segment and 1–8-character letter/digit segments; it does not check registered languages or complete BCP 47 validity.
- Accept-Encoding
- Content codings such as gzip, br or identity, each with an optional q value.
- Offered representations
- One per line as "media type | language | encoding | label". Use "-" for a missing language or encoding. Up to 30 lines.
- Equal-score policy
- Real servers differ when two representations score the same. The model can either follow the declared order or refuse to name a single winner.
Supported inputs and limits
- Up to 24 ranges per supported preference field and 30 offered representations.
- Selected media/language/encoding subset only; no request, codec test, server configuration or guarantee of actual origin selection.
- Blank fields model absent headers.
- Quoted delimiters and repeated quality parameters are refused.
- Preference scores and exclusion reasons use the documented tie policy.
More details
No live server response or complete implementation of every negotiation variant is claimed.
Language syntax uses 1–8-letter first segments and 1–8-character letter/digit segments.
Registered language names, complete BCP 47 validity and extended matching are outside this model.
Where your input is processed
This tool does not upload your inputs. See the limits for anything saved on your device.
The ranking order is a declared model
After determining each matching quality, the model sorts acceptable choices by media quality first, then language quality, then encoding quality. It does not multiply the three scores. A zero in any applicable dimension excludes that representation. Equal triples use listed order or remain unresolved, according to your choice. HTTP does not require every server to use this selection algorithm; check your application’s actual policy.
Questions about HTTP Content Negotiation Tester
Will a real server choose this winner?
It may use another selection or tie policy.
Does q=0 mean a low preference?
An applicable zero-quality value excludes that modeled choice.
Does it compress the offered data?
No. Encoding labels describe supplied representations; no payload is transformed.