Skip to main content
Experimental features are functional but still being validated on production infrastructure. Test on staging first, and understand what you’re turning on before you turn it on. Guest Mode in particular changes what certain visitors see, and that’s not a decision to make casually.

Guest Mode

Guest Mode shows PageSpeed-style testing tools (PageSpeed Insights, GTmetrix, and similar synthetic testing bots) a more aggressively optimized version of your page, while real human visitors continue to see your full site. It exists for one specific, narrow situation: you’ve applied every real optimization available (everything on the Optimization and Images pages, a solid cache setup), and you’re still short of a target score for a reason that’s cosmetic to the test rather than a real problem for visitors.
Real optimizations benefit actual users. Guest Mode benefits a score. Those aren’t the same thing, and the order you reach for them matters: exhaust the real levers first. If you find yourself reaching for Guest Mode as your first move instead of your last, that’s worth reconsidering.

Levels

Guest Mode custom CSS

A separate CSS field, appended to the combined stylesheet only when Guest Mode is active, meaning only PageSpeed-style tools ever see it. Its practical use is patching layout breakage caused by stripping JavaScript for the test: animation libraries like GSAP often leave elements stuck at opacity: 0 waiting for JS that Guest Mode just removed. Use browser DevTools against a Guest Mode render to find elements stuck invisible, then fix them here with something like opacity: 1 !important or visibility: visible !important.

Guest Mode Cache Isolation

A real, separate problem this section solves: without isolation, if a testing bot happens to be the first visitor after a cache flush, Batcache or Edge Cache can store that stripped-down bot render and then serve it to actual human visitors. Isolation keeps guest (bot) responses out of the shared cache pool entirely, so this never happens. If isolation is on, turn on warming the Guest bucket too, otherwise the bot-facing cache stays cold and the first real bot visit after a flush pays the full render cost anyway.
Isolation only does anything while Guest Mode itself is enabled above. With Guest Mode off, there’s no divergent render to isolate, so the isolation toggle has no effect and the screen tells you so directly.

Layer 1: plugin-level isolation

A single checkbox. Sends no-store headers and sets DONOTCACHEPAGE for guest (bot) visitors, which prevents guest renders from being stored in Batcache or Edge Cache in the first place.

Layer 2: wp-config.php bucket separation

Batcache serves cached pages from advanced-cache.php before any plugin loads, which means Layer 1 alone can’t fully separate bot and human traffic at serve time, only at store time. This layer adds a snippet to wp-config.php (above the “That’s all, stop editing!” line) that keys the Batcache bucket itself on visitor class, so bot and human requests never share a cache entry at all, not even briefly. The screen shows whether the snippet is currently detected in wp-config.php and provides the exact snippet to paste in, click the code box to select it all. After adding it, run Flush All Caches once so any existing mixed cache entries (from before the snippet existed) get cleared out.
If wp-config.php isn’t writable and you change your guest-mode level or isolation setting, Aero can’t remove the wp-config snippet automatically. It flushes all caches regardless and tells you plainly that the snippet needs manual removal, and that until then bots remain keyed into a separate bucket even if you’ve turned isolation off elsewhere. Don’t assume a setting change fully took effect without checking this notice.

A candid take on when to use this

Guest Mode optimizes for a synthetic lab test. That’s the honest framing, and it matters because a passing score there doesn’t touch what a real visitor experiences or how Google actually ranks you. Google’s page-experience signals in Search come substantially from CrUX (Chrome User Experience Report) field data, collected from real Chrome users out in the world, not from Lighthouse or PageSpeed Insights lab scores. A Guest Mode setup that impresses a PSI test can leave your real CrUX numbers, and your real visitors’ experience, completely unchanged. If your goal is genuinely better Core Web Vitals and better rankings, the Optimization and Images pages are where that work actually happens. Guest Mode is for the narrower, legitimate case of a client, stakeholder, or agency contract that specifically checks a PageSpeed number as a deliverable, and even then, Basic over Extreme, and only after the real work is done.