Edge Copilot Can Now Read Every Tab You Have Open

On 13 May 2026, Microsoft pushed a significant update to Edge. Copilot can now read across every tab you have open and answer questions about all of them at once. It also remembers your previous conversations, can access your browsing history, and the whole package is now on mobile. Here’s what actually changed — and what you’re agreeing to when you switch it on.

What tab-reading actually does

Until now, Copilot in Edge could see the page you were on and answer questions about it. Everything else in your tab bar was out of bounds. The May 2026 update removes that boundary.

With permission granted, Copilot ingests the content of every open tab simultaneously and treats them as a single working surface. You can ask it to compare prices across five retailer tabs, pull a specific spec from an article you opened an hour ago, or summarise a stack of research without switching between anything. No copying, no pasting, no tab hopping. Ask the question; get the answer.

Microsoft’s announcement frames this around mundane use cases — planning a trip, comparing restaurants. The framing undersells it. A sidebar chatbot that waits to be told what to read is one thing. An assistant that has already read everything you have open is a genuinely different tool.

No setup is required. Click the Copilot icon in the top-right corner of Edge, grant permission when prompted, and it starts. The permission is explicit — you see exactly what you’re enabling before anything happens, and Edge displays clear visual indicators while Copilot is active.

Everything else in the May 2026 update

The tab feature is the headline, but the update is broader. These are the other changes worth knowing about:

• Browsing history context. Copilot can optionally access your browsing history to inform its answers — surfacing something relevant from a page you visited last week, for example. Requires a separate, explicit permission opt-in.
• Long-term memory. Copilot now retains context from previous conversations across sessions, on both desktop and mobile. Ask it something today; it has context from the week before. Microsoft says this is designed to make responses progressively more relevant over time.
• Study and Learn mode. Copilot can convert an article you’re reading into a structured study session or interactive quiz. Genuinely useful for anyone doing technical reading or revision.
• AI-generated podcasts. Copilot can turn web content into a short audio summary. Whether you’ll use it depends entirely on your workflow — but it’s there.
• Journeys, now on mobile. Journeys organises your browsing history into topic clusters with summaries and suggested next steps. It’s been on desktop for a while; the Edge mobile app gets it for the first time with this update.
• Vision and Voice, expanded to mobile. Share your screen with Copilot and talk through what you’re looking at, hands-free. Previously desktop-only; now available in the mobile app as well.
• Redesigned new tab page. Chat, search, web navigation, and Journeys are now combined in a single redesigned new tab page. More opinionated than before.

Copilot Mode is gone — and the split matters

Copilot Mode — Microsoft’s previous AI-native browsing experience in Edge — is being retired with this update. It did similar things to what’s described above, including tab reading, but it also had agentic capabilities: completing bookings, filling forms, taking actions on your behalf while you stepped back.

Those agentic features haven’t been removed. They’ve been separated into a standalone tool called Browse with Copilot. The distinction Microsoft has drawn is between contextual reading (passive, helps you decide) and autonomous action (active, does things). Keeping them in separate, separately-governed experiences is a cleaner position — for compliance, for user trust, and for the practical reason that most people want very different levels of scrutiny applied to each.

For most users, this is the right architecture. An assistant passively reading everything you have open is useful and relatively low-stakes. Autonomous action is something you want to consciously invoke, not have ticking away in the background.

What you’re actually agreeing to

Three distinct permissions are in play here, and they’re independent of each other.

Tab reading is triggered when you click the Copilot icon and grant access. It applies to that session. Edge shows visual cues while Copilot is reading — you’re not left guessing whether it’s active.

Browsing history access is a separate opt-in. It’s not bundled with tab reading. The benefit is more contextually aware responses over time; the trade-off is that a broader slice of your browsing data flows through the service.

Long-term memory is opt-in on both desktop and mobile. Previous conversations are stored and inform future sessions. Microsoft hasn’t published a detailed breakdown of where that data lives or for how long.

The outstanding question for corporate and regulated environments is whether tab content is processed locally or transmitted to cloud infrastructure. Microsoft hasn’t clarified the on-device versus cloud split specifically for tab reading. That’s a gap worth resolving before enabling it across professional sessions where sensitive documents or client data might be in open tabs. IT administrators can govern which Copilot features are available to users via Edge Group Policy — the full policy reference is on Microsoft Learn.

Who gets the most from this

Cross-tab reading earns its keep when you’re genuinely juggling information across multiple sources — comparing products, reconciling specs across documentation pages, or doing research where the answer lives across three different articles. The cognitive overhead of tab switching is real, and removing it matters.

The long-term memory feature is the more interesting proposition over time. An assistant that accumulates context about what you’ve been working on — across sessions, not just within one — starts to behave more like a genuine research collaborator than a chatbot you have to re-brief every time. Whether Microsoft executes that well remains to be seen. The architecture is right.

For casual browsing with a handful of tabs, none of this changes much. But if you routinely have twenty tabs open and spend real time hunting between them, this update directly addresses that friction.