ChatGPT can read and control Chrome tabs, on every plan

On October 15, around 1:02 PM PT, ChatGPT posted that the Chrome extension can now use the tabs you already have open. In Chat, it can look across those tabs and gather info. In Work, it can control them: send messages, fill out forms, download files, and group a pile of tabs. A follow-up post says this is available for all ChatGPT plans, and links the Chrome Web Store listing.

The Verge reports that it works by tagging Chrome in the web or desktop app, and that the sidebar gets your ChatGPT history so you can keep working on the page you are viewing. One of OpenAI's sample prompts, quoted there, is to fill a registration form from an event brief and stop before submitting.

The extension doc is broader than the Chrome post. It says Chrome, Edge, Brave, Opera, and Vivaldi can do tab mentions and browser control from the desktop app. Side chat is on Chrome, Edge, Brave, and Vivaldi. Opera does not get side chat. Setup runs through Settings, then Computer Use, in the desktop app.

What it costs

No new price. The post says every ChatGPT plan. The extension's Chrome permission prompt, listed in that doc, is broad: access the page debugger, read and change all your data on all websites, read and change browsing history on signed-in devices, notifications, bookmarks, downloads, native applications, and tab groups. ChatGPT still asks before it uses a new site, and the same settings hold an allowlist and a blocklist. "Allow for all sites" turns those prompts off. The doc says OpenAI does not keep a separate full record of browser actions. What it stores is whatever enters the chat. Your ChatGPT data controls apply to that.

Should you switch

Try it on tabs you would not mind a model reading. Leave the per-site prompt on, and keep the habit in OpenAI's own example: stop before submitting. Do not turn on "Allow for all sites" for a first run.