A cacheable resource fetched twice comes from cache the second time
The second fetch of the resource is served from the HTTP cache. N/A under a controlling service worker, and when Resource Timing does not record both fetches (severity: warn).
What a detector infers
A browser that is allowed to cache a response does. The probe fetches an immutable, cacheable resource unique to this visit twice, then reads Resource Timing for the second fetch: an ordinary browser serves it from its HTTP cache, which shows as a transfer size of 0 or a cache delivery. Request interception through the DevTools protocol changes that.
Playwright turns interception on when a script uses route(), when it sets httpCredentials, and when its proxy needs credentials — and with interception on, Chrome stops serving from the HTTP cache, so both fetches go to the network. Measured by counting requests at the server: Chrome and Edge on Windows, Linux Chrome and plain Playwright all sent one; Playwright with route() and Playwright with httpCredentials sent two.
Interception is set up below the page, so nothing a script removes from the page restores the cache. Two honest exceptions: DevTools' own 'Disable cache' box does the same thing for a developer, and a service worker in front of the request answers from its own cache, so a controlling service worker makes the row N/A. It differs from request-cache-headers, which reads the headers on the first request; this row reads whether the second one went to the network at all.
How to resolve it
Do not route requests through the automation client: drop route() handlers and httpCredentials, and give a proxy its credentials at the network layer — an allow-listed IP or a local forwarder that adds them — instead of answering the proxy's challenge through interception.
Nearby checks in HTTP request vs the runtime
- Client hints on the wire carry the values JavaScript reports
client-hints-wire-vs-jsA server that asks for client hints with Accept-CH receives, on the next request, headers that describe facts JavaScript can also read… - The image Accept header is the one the claimed Chrome major sends
image-accept-vs-claimed-majorThe Accept header Chrome sends for an <img> is compiled from the image formats the build decodes by default, and it changed at 155: JPEG XL… - A WebSocket handshake carries the no-cache pair every browser sends
websocket-handshake-no-cacheA WebSocket connection starts as an HTTP request that is never cacheable, and browsers mark it that way. - Accept-Language leads with the same language as navigator.languages
accept-language-vs-navigatorAccept-Language and navigator.languages are two views of one browser preference: the header is serialised by the network stack, the array is… - Global Privacy Control reads the same in the page, in a worker and on the wire
gpc-header-vs-navigatorGlobal Privacy Control has three faces: navigator.globalPrivacyControl in the window, the same property in workers, and the Sec-GPC: 1… - The client-hint headers match navigator.userAgentData
sec-ch-ua-vs-uadataUser-Agent Client Hints exist in two places at once, which is what makes them checkable. - Sec-Fetch-* describes the request this page actually made
sec-fetch-metadataSec-Fetch-Site, Sec-Fetch-Mode and Sec-Fetch-Dest are attached by the browser to describe the request's own provenance — where it came from… - The order your browser emits its HTTP headers in
http-header-order-readoutA browser emits its request headers in a fixed sequence compiled into its network stack, and that sequence differs between engines…
Clearcote is a browser built for fingerprint coherence
It is a Chromium fork, maintained by the same people who wrote this reference. It ships as a compiled browser rather than as a stealth script injected into someone else's — which is a description of how it is built, and is not an argument about how it behaves on this check.
This audit takes no position on how Clearcote scores on HTTP request vs the runtime checks, on this one, or anywhere else. It has no baseline corpus of other people's fingerprints to rank you against and no vendor scoreboard — nearly every check is self-referential, asking one browser the same question through two independent APIs and reporting whether both answers can be true at once. It runs identically on any browser, including ours. Run it on yours and read the result yourself.
See the other checks in HTTP request vs the runtime — the family http-cache-honoured belongs to.
