A desk facing a walkway gets read by dozens of short, distant glances a day. Docognito removes the one cue that gives ChatGPT or Claude away at that range, the alternating message bubbles and repeating assistant name, and redraws the page as plain document paragraphs instead. A stop-and-read up close can still notice the question-and-answer pattern underneath.
Why distance changes what gives you away
A neighbour sitting beside you reads words. Somebody forty feet down a corridor, moving at walking pace, reads shapes and colour. They are not close enough to make out a sentence, but they are close enough to register a narrow centred column, two alternating background tones, and a name that repeats every few lines. That combination is the single most recognisable visual signature a chat interface has, and it is legible at a distance that no amount of small text or dim brightness fixes.
This is why a desk facing a walkway is a harder problem than a desk against a wall. It is not one glance from one direction. It is dozens of glances a day, from anyone who happens to walk that route, and every one of them is reading the shape of the page rather than its content.
A second monitor angled toward the same corridor makes the problem worse again, because it doubles the surface reading the same shape at the same distance. Repositioning a monitor away from a walkway helps on its own, and it is worth doing regardless of any software change, but a desk in a fixed layout, a shared bank of workstations, or a room where the monitor cannot move away from the door still needs the page itself to stop announcing what it is before distance and angle can do the rest of the work.
What the disguise actually removes
Docognito redraws the visible page, redrawn as a document editor as a plain document. Message bubbles become paragraphs on a page. The alternating background colours are gone. The assistant name that used to repeat down the left margin of every reply is gone too, because there is no assistant name in a document. the browser tab title and the tab icon change alongside it, which matters for anyone glancing at your taskbar rather than the monitor itself.
None of that requires you to do anything mid-conversation. It is Shift and Escape, which both hides the disguise and brings it back, so the moment you hear footsteps, or simply want the screen calmer by default, one keypress handles the whole page at once.
A distant glance versus a close read
| Distance | What they actually notice |
|---|---|
| A walk-past, several seconds | Layout and colour, not words |
| A pause at your desk, a minute or more | Individual sentences, and their order |
| A deliberate read, standing still | The pattern of a question and a considered answer |
The disguise is built for the first row, and it works well there, because a distant glance was never going to read your words in the first place. It does less for the third row, and it is worth being honest about that rather than pretending otherwise.
What a close, deliberate read would still notice
the visible page, redrawn as a document editor renders replies as ordinary paragraphs, so from across a room it reads as text on a page, which is exactly the goal. But somebody who actually stops and reads it, rather than walking past, can still notice the rhythm underneath the page: a short block of text, then a longer, more structured block that answers it, repeated down the document in the same order every time. An ordinary document does not usually read like a transcript of questions and answers. A patient reader can pick up on that even with every bubble and every name gone.
That is a different threat from the one this article opened with. A walkway is mostly short glances at a distance, and the disguise is built exactly for that. A colleague who pulls up a chair and reads along is a closer, longer look, and no visual redraw changes what the words underneath actually say.
It is worth naming that difference plainly rather than letting the page's document styling do more persuading than it has earned. A person who reads carefully is testing content, not shape, and content is outside anything a visual disguise was ever built to change.
Reduce the close-read risk too
Position the monitor, not just the software
A screen angled even ten degrees away from a corridor cuts the readable range dramatically, and it costs nothing to adjust once.
Turn the disguise on by default, not as a reaction
Waiting until you hear footsteps means the page is exposed for the seconds it takes to notice someone. Leaving it on by default removes that gap entirely.
Treat a paused visitor differently from a walking one
If someone stops rather than passes, the shape of the page is no longer doing the work. Switch to a different task, or accept that a close read will notice the pattern, before assuming the disguise alone covers it.
What distance hides and what a closer look still finds
For a desk that faces a route people walk past all day, Docognito solves the part that distance actually exposes, the recognisable shape of a chat window, and it does that with a single keypress and no change to how you work. If your problem is a colleague who stops to read, combine it with where your monitor points, because no redraw of the page changes what a patient reader can still work out.
Add Docognito and turn it on before the first walk-past of the day, not after.