Turn Docognito on before you join, and ChatGPT or Claude renders as a document for the call. Shift plus Escape flips it back to the real interface the moment someone asks to look, and the same shortcut brings the disguise straight back once they do.
The workflow in three moves
The whole approach comes down to timing the one toggle correctly and knowing the one shortcut. Everything else about using ChatGPT or Claude on a call stays exactly as it normally would.
Install the extension once
Add Docognito to Chrome and open claude.ai or chatgpt.com in a fresh tab, since a tab open before install will not pick it up until reloaded.
Toggle it on before you join the call
Switch the disguise on while setting up, not after you are already on camera or already sharing. The page becomes a document editor immediately.
Know the shortcut before you need it
Shift and Escape, which both hides the disguise and brings it back, and it works instantly the moment the extension is active in that tab.
What "renders as a document" actually looks like
Once the disguise is on, the chat interface disappears and a page styled like a document editor takes its place: a title bar, a menu row, a page layout with margins, and a typing area that looks like a cursor sitting in a paragraph rather than a message composer. the visible page, redrawn as a document editor, and typing, sending and reading a reply all continue to work exactly as before, just inside that different shell.
Nothing about the underlying conversation changes. The words you type still go to ChatGPT or Claude the same way they always did, and the reply still comes back the same way. Only the shell around it is different.
To someone glancing at your screen, none of this reads as unusual. A page with a title bar, a menu row and a block of body text sitting inside margins is the visual grammar of ordinary office work, the same general shape as most other tabs open across a normal working day.
Using the shortcut mid-call without it looking odd
If someone asks to see your screen, or leans over to look, pressing Shift plus Escape switches the page back to the real interface in the same instant. Once they have looked and moved on, pressing it again brings the document view straight back.
The best moment to use it is a natural pause, not mid-sentence, and once is better than several times in a row. A page that changes shape once reads as unremarkable. A page that keeps flickering between two very different layouts reads as exactly the kind of thing worth a second look.
The switch itself is visible for a moment on a live screen share
Toggling the shortcut while someone is actively watching your shared screen shows as a brief, visible change from one layout to another. It is fast, but it is not invisible, so the better habit is turning the disguise on before the call rather than reaching for the shortcut on screen.
Pairing this with how you share your screen
The disguise handles the page. It does nothing about which sharing mode sends the browser frame around that page along with it, and that half of the problem lives in whichever call platform you are using rather than in the extension itself. A tab share in Google Meet, or a window share in Teams, removes the address bar from a live view the same way Docognito removes the page's own look, and the two are separate changes rather than one change wearing two names.
Neither one alone finishes the job. Skip the disguise, and the shared page still looks like a chatbot even inside the narrowest sharing mode available. Skip the narrow sharing mode, and the disguised page still arrives with the address bar attached, because the browser frame is not something a page level change can ever reach.
What this workflow does not cover
Doing every step right still leaves the address bar showing
the address bar, which still reads claude.ai or chatgpt.com, and Docognito never touches that, on purpose, because rewriting the URL to hide it broke the site underneath when it was tried. Someone who reads the address bar closely, rather than glancing at the page, learns what site you are on regardless of the disguise.
Toggling off again after the call is worth building into the habit too, if only so the page returns to normal for whatever you do next in that tab. This is a separate question from what stays on your computer once the call ends: the disguise governs only what a page shows while it is open, not what your browser remembers about having visited it at all.
Timing the toggle before the call starts
Turn Docognito on before you join, not during the call, and keep Shift and Escape, which both hides the disguise and brings it back in mind for the moment someone asks to look at your screen directly. That covers the page, the tab title and the icon, which is what a glance actually takes in. The address bar stays as it is the entire time, and no step in this workflow changes that.