OpenAI’s ChatGPT Chrome extension can now read the browser tabs you already have open, and the sidebar has been upgraded so that ChatGPT - and your conversation history - sits next to whatever page you are on. The Verge reported both changes on October 15, 2026.
Key facts
| Fact | Detail |
|---|---|
| Question answered | ChatGPT Can Read Your Open Chrome Tabs Now |
| Short answer | OpenAI’s ChatGPT Chrome extension can now read the browser tabs you already have open, and the sidebar now carries full ChatGPT plus your conversation history. |
| Reading time | ~3 min |
| Last updated | 2026-10-16 |
What actually changed
The report describes two upgrades in one release:
- Tab awareness. With the extension installed, ChatGPT can view your open browser tabs when you tag Chrome in ChatGPT’s web or desktop app. OpenAI posted a demo video of someone asking ChatGPT to compare three open tabs - the sofas they were shopping for.
- A fuller sidebar. The sidebar now carries what The Verge calls “the full power of ChatGPT” plus your conversation history in Chat or Work, without leaving the page you are reading.
There is a third piece that moves past reading. The report says ChatGPT Work can control your Chrome tabs for jobs like sending messages, filling out forms or downloading files, and that it can also organize or group open tabs. The closest comparison in the article is Google’s Gemini-powered “auto browse” feature.
The Verge also notes that OpenAI’s ChatGPT Atlas browser is gone - so the extension, not a separate browser, is where the browser-side capability now lives.
Read the full piece here: https://www.theverge.com/ai-artificial-intelligence/1012510/openai-chatgpt-google-chrome-tabs
Why this matters if you run an agent on Windows
This is a clean illustration of a split that has been forming all year. Where an agent lives decides what it can touch:
- A browser extension sees tabs, pages and forms. Its reach stops at the edge of the browser.
- A desktop agent sees the machine. Files, folders, installers, command-line tools, local builds - the work that never routes through a tab.
For a Windows user running a local agent such as MOOGH, the practical question is not “which one wins” but “which one holds the context I need”. A task like “compare these three product pages” is browser work. A task like “rename and sort this download folder, then run the build” is desktop work, and no tab context helps with it.
The other thing worth noticing is the direction of travel: reading a page is an easy sell, and acting on it - sending, filling, downloading - is a permission question. If you are still deciding where you want an assistant to have hands, start with [download MOOGH] and keep the acting side on your own machine, where every step is approved and logged.
What to check before you rely on it
- Did you actually tag Chrome? The tab-reading behaviour is described as triggered from the conversation, not as background monitoring.
- Which surface are you in? The article distinguishes the web app, the desktop app and the sidebar extension.
- Who is signed in? The sidebar upgrade is tied to your ChatGPT conversation history in Chat or Work.
- What is the extension allowed to act on? Reading tabs and controlling tabs are different permissions in practice, even when they ship together.
FAQ
Is tab reading on by default? The report describes it as a capability of the upgrade rather than a background behaviour: you raise Chrome inside the conversation.
Which tabs can it see? Your open tabs, per the article’s wording, with the demo showing a comparison across three tabs.
Does this work outside Chrome? The update is described as shipping with the Google Chrome extension. Other browsers are not covered in the report.
Should I uninstall a desktop agent for this? No. They cover different ground. The extension covers what is in the browser; a desktop agent such as MOOGH covers the local files, commands and builds a browser cannot reach.
Related reading: