A changed tab title and icon convinces a glance, not a search. Eyes register shape and colour the moment they cross a screen, well before they would stop to read a label, so a document icon and a plausible name give a glance nothing worth a second look. Nothing about the change survives someone who actually stops to look closely.
What two seconds actually resolves
A glance across a desk, or a look thrown sideways in a meeting, does not involve reading in any real sense. Reading a word takes a fixation, a brief stop where the eye holds still long enough to resolve individual letters, and a passing glance rarely affords even one. What it does afford is a much cruder judgement: does this look like the kind of thing expected in this spot. A monitor covered in paragraphs and a plain toolbar reads as a document at that level of judgement, without a single word being read. A monitor showing a chat bubble and a coloured logo reads as something else at the exact same level, again without anyone reading anything at all. The disguise is not competing against careful reading. It is competing against a snap categorisation that finishes before reading would even begin.
Why shape and colour win before words do
Vision does not process everything on a screen at once, or at the same speed. Fine detail, the kind needed to tell one letter from another, is only available in a narrow part of the visual field, the part pointed directly at whatever the eye is fixating on. Shape and colour are read far more broadly, including in parts of the visual field that never come into sharp focus at all. That is why a coloured icon in a tab strip registers before the label beside it does, and often before the eye has moved to that tab at all. Colour and rough shape get picked up in passing, while the word next to them needs a direct look that a glance frequently never grants. A tab redrawn with a plain page icon and a document coloured favicon is judged on exactly the cues a glance is equipped to use quickly, while the label sitting beside it is almost a formality by comparison.
Why a plausible label beats a blank one
Not every disguise is equally invisible, and the difference matters more than it looks. A tab with no title, or one replaced by an obviously placeholder string, sits out of place next to a row of ordinarily named tabs, and an item that looks out of place is exactly what draws a second glance, even from someone not looking for anything in particular. An unusual gap in an otherwise normal pattern is its own kind of signal. A tab that instead reads as an ordinary document, the kind of title and icon seen a hundred times before, blends into that same row rather than interrupting it. The goal is not to remove information from the tab. It is to replace one recognisable pattern with a different, equally ordinary one, so nothing about the row invites a second look in the first place.
Where this kind of change is worth having
Most of the exposure people describe is exactly this kind of glance: a coworker walking past an open desk, a family member passing behind a laptop at a kitchen table, a classmate glancing sideways during a lecture. Nobody in those situations is stopping to read a tab carefully or checking which icon set is installed. They are looking briefly at whatever a screen shows and moving on, which is precisely the layer a changed tab title and icon operates on. Docognito applies both automatically the moment a claude.ai or chatgpt.com tab loads, and the mechanics of how it keeps winning that fight are covered on the how to page for changing the tab title and icon. What matters here is narrower than the mechanics: whether the two seconds someone's eyes cross a screen registers anything worth a second look, and a plausible document tab gives that glance nothing to catch on.
What this does not cover
A changed tab title and icon is a glance level change only, and it says nothing about a deliberate look. What survives that kind of closer inspection, the developer tools and the address bar included, is covered separately on the page about hiding the address bar. Anyone reading that far past a glance has already stopped doing the thing this page is about.
A tab label answers a glance, not a close inspection
A glance and a deliberate, close inspection call for different answers here. When the risk in front of you is a glance, a changed tab title and icon genuinely answer it, because a glance is a categorisation problem, not a reading problem, and Docognito changes exactly the cues that categorisation relies on. When the risk is a deliberate, close inspection of your screen, no tab label reaches that, and treating a glance level fix as though it does only manufactures a false sense of cover. Knowing which of the two situations is actually in front of you is most of the decision.