Features
Docognito redraws ChatGPT and Claude as a plain document editor. Typing, sending, and the reply streaming in behave exactly as they did before; only the page around them looks different.
Everything the disguise changes
Six things make up the whole feature set. None of them requires switching anything off first, and none of them is held back for a later paid version.
Document mode on both sites
The page becomes a plain document view on ChatGPT and Claude, with no chat bubbles, no assistant avatar, and no product branding left on screen. Whichever of the two you are using, the disguise matches it.
One shortcut, both directions
The shortcut is Shift and Escape, which both hides the disguise and brings it back. Press it once and the real page shows underneath; press it again and the document view returns. There is one combination to remember, not a separate toggle for hiding and for restoring it.
The tab title and the icon change too
Docognito changes the browser tab title and the tab icon as well as the page itself, so an open tab in the tab strip reads and looks like an ordinary document rather than an AI chat.
Uploads work inside the disguise
Dropping a file, pasting an image, or attaching a document from the file picker all work while the document view is showing. There is no need to switch the disguise off first, and a small chip confirms once the file has actually reached the assistant.
Switching models without leaving the disguise
A model chip reachable from the document view opens each assistant's own current model list, so choosing between the models your plan offers does not mean dropping back to the real interface first.
Dark mode, if you want it
Dark mode is a setting in the extension's own popup, not a menu inside the disguise. Neither ChatGPT nor Claude's own web app changes; only Docognito's document view does, on demand.
Typing, sending, and replies stay exactly as they were
Everything typed goes into the real composer for whichever assistant is open. Selecting text, deleting it, using the arrow keys, and pasting all behave the same way they did before the document view was switched on.
Sending a message and watching the reply stream in also behave the same way. Nothing is retyped, resent, or delayed by the page drawn around it; the disguise changes what is on screen, not how the conversation actually works. Add Docognito and it is the same on your next tab.
What it does not do
A visual disguise changes what a person looking at the screen sees. It does not change what the browser, the network, or the assistant on the other end does with the conversation. Read this list before relying on Docognito for something it does not cover.
What stays exactly as it was
- Docognito does not change the address bar, which still reads claude.ai or chatgpt.com.
- Docognito does not change your browser history.
- Docognito does not change anything a workplace network, proxy, DNS log or monitoring agent can see.
- Docognito does not change anything a managed or employer-administered device can see.
- Docognito does not change the conversation itself, which OpenAI and Anthropic still receive and store.
- Docognito does not change anything captured by screen recording or proctoring software.
- Docognito does not change the page source, which anyone opening developer tools can read.
Common questions
Does Docognito work on both ChatGPT and Claude?
Yes. It runs on ChatGPT and Claude, covering claude.ai, chatgpt.com, and the older chat.openai.com address. Gemini is not supported.
Is there a separate shortcut for each site?
No. Shift and Escape is the only shortcut, and it works the same way on both sites: press it once to hide the document view, press it again to bring it back.
Does turning on dark mode change the real ChatGPT or Claude interface?
No. The dark mode setting lives in the extension popup and only affects Docognito's own document view. Neither assistant's underlying web app is touched.
Do file uploads and the model picker still work while the page looks like a document?
Yes. Both are reachable from inside the document view. Uploading does not require switching the disguise off, and the model picker reads each assistant's own current list.
Add Docognito to Chrome
Free, one permission, and the document view is a keypress away on your next tab.
Add to Chrome, it's free