ChatGPT can now work with pages already open in Chrome. OpenAI says its upgraded Chrome extension lets people tag Chrome in ChatGPT’s web or desktop app so the assistant can view those tabs and use their contents in a request.
Ask ChatGPT to compare what you have open
The feature is aimed at tasks that involve information spread across several pages. In a demo video, someone asks ChatGPT to compare three open tabs containing sofas they are considering. That kind of request turns a collection of separate browser pages into material for one comparison.
OpenAI’s examples show other ways to use the open tabs. A user could ask ChatGPT to pull vendors, pricing, key features, and source links into a comparison sheet, or review contractor services and draft an email to a suitable provider. The pages supply the information for the task; the user describes what they want ChatGPT to do with it.
The extension can also help with a form that is already open. One example asks it to use an event brief to fill out a registration form and stop before submitting. Another asks it to fill out a form ahead of a dentist visit and add the appointment to work and personal calendars.
The sidebar brings ChatGPT into the page
The update also changes ChatGPT’s Chrome sidebar. OpenAI describes it as providing “the full power of ChatGPT” and access to a user’s entire conversation history in Chat or Work, without leaving the page they are viewing.
That setup connects the page in front of the user with conversations they have already had in ChatGPT. Someone can keep a browser page visible while using the sidebar, rather than switching away from it to continue a conversation. The extension therefore adds both a way to use open tabs as context and a way to keep ChatGPT alongside the browsing session.
ChatGPT Work can take actions in Chrome
For ChatGPT Work, the extension goes beyond reading and comparing pages. OpenAI says it can “control” Chrome tabs for tasks such as sending messages, filling out forms, or downloading files. It can also organize or group open tabs.
Those capabilities change the role of the assistant: a request can involve gathering information from pages and then carrying out a browser task. The sample instruction to fill a form but stop before submitting also illustrates how a request can set a boundary around an action. Users can make that boundary explicit in the task they give ChatGPT.
A browser extension built around existing tabs
The update arrives after OpenAI’s ChatGPT Atlas browser “is gone,” according to the report. The Chrome extension instead works with the browser tabs a person already has open. That distinction is central to the feature: people can ask ChatGPT to work across current pages without moving the task into a separate browser.
In practice, the examples span research, comparison, drafting, form filling, and tab organization. The extension’s usefulness depends on having relevant pages open and describing the task clearly. For requests involving consequential actions, the examples show that users can specify what the assistant should do and where it should stop.