In short. Edge can provide Copilot with URLs, page titles, page content or screenshots, and browsing history as prompt-dependent context.
Definition: In Microsoft Edge, browsing context is information from the browser that Copilot may use with a prompt. Microsoft lists the URL, page title, page content or screenshots, and browsing history as possible context clues.
Key facts
- Microsoft's Context Clues help lists four browsing-context types: URL, page title, page content or screenshots, and browsing history. It places the setting under More menu (…) > Copilot and AI > Allow Copilot to read context clues on the web.
- Edge selects sources according to the prompt. A request about page content can use the current webpage, page title, open tabs, and browser history, while a prompt without a page reference uses the URL, page title, prompt, and previous conversation history, according to Microsoft's Edge guide. The same guide says page titles and URLs from open tabs can supply suggested prompts.
- Browse with Copilot is a separate action-taking mode. Its Microsoft help page says screenshots taken during an action are saved with the conversation, are not used for training, and are retained for up to 30 days unless the conversation is deleted sooner.
Which data can Edge provide?
Microsoft's Context Clues documentation names the data classes directly:
- URL
- Page title
- Page content or screenshots
- Browsing history
The Edge guide adds that the current webpage, page title, open tabs, and browser history can be used for a prompt that refers to page content. For a request such as trip planning that does not refer to page content, the same guide says Edge uses the URL, page title, prompt, and previous conversation history. The source therefore describes prompt-dependent selection rather than a fixed payload for every request.
The page title and URL of the current page can also support suggested prompts when Copilot opens. The Context Clues help says those suggestions do not use Context Clues, and that generic suggestions appear when the setting is off.
What does the setting control?
The relevant Edge setting is:
More menu (…) > Settings > Copilot and AI >
Allow Copilot to read context clues on the web
Microsoft's Context Clues help says browsing context is controlled by this setting. When it is off, Copilot has no access to browsing data and will not answer questions or perform requests based on the websites or browsing data. The Edge guide describes the same control as preventing use of the current webpage, page title, open tabs, and browser history in responses.
Microsoft also says that browsing context is not saved in the conversation history for Context Clues. The prompt and Copilot response are saved. This storage statement applies to the Context Clues workflow described on that page.
How do screenshots and history differ in Browse with Copilot?
The separate Browse with Copilot help page describes screenshots captured while Copilot performs an action. It limits them to the tabs Copilot is working on in the current browser window and says they are saved with the conversation so the work can be reviewed. At the October 2026 review, Microsoft's page stated that deleting the conversation deletes its associated screenshots; otherwise, screenshots are retained for up to 30 days.
That page also says text prompts, Copilot responses, and a record of the work are saved in conversation history. Information entered directly into a webpage, such as personal data in a form, is not saved by Copilot, although Edge may save it. These rules describe Browse with Copilot's action history, while the Context Clues help lists screenshots as a possible context type without giving a separate capture or retention schedule.
Related terms
- Microsoft Copilot
- Which Edge setting permits Copilot to use the current webpage?
- Is Edge browsing context saved in Copilot conversation history?
- How long are Browse with Copilot screenshots retained?
FAQ
Does Edge always send the full page to Copilot?
Microsoft documents prompt-dependent source selection. A prompt about page content can use page information, open tabs, and browser history, while a prompt without a page reference uses a narrower set that includes the URL and page title.
Can browsing history be used as Copilot context?
Yes. Microsoft lists browsing history as a browsing-context type and says the browser history can be used for prompts that refer to page content.
Can the setting stop browsing context use?
Yes. Turn off Allow Copilot to read context clues on the web in Edge under More menu (…) > Settings > Copilot and AI. Microsoft says this removes Copilot's access to browsing data for answers and requests based on websites or browsing data.
Are screenshots always kept in conversation history?
The Context Clues page says browsing context is not saved in conversation history. Browse with Copilot has a separate rule: action screenshots are saved with that conversation and retained for up to 30 days unless the conversation is deleted sooner.
Sources
- Microsoft Support, "How Context Clues work in Copilot in Edge", undated
- Microsoft Support, "Getting started with Copilot in Microsoft Edge", undated
- Microsoft Support, "Browse with Copilot", undated
Check how a page reads to answer engines. Open the GEO score analyzer