Microsoft Copilot Vision is a new way to have AI work with you as you use your Windows PC or Mac, browse the web in Microsoft Edge, or view the world from your phone. It acts like a second set of eyes that can scan, analyze, and offer insights based on what Copilot sees. When you start a Copilot Vision session, Copilot can view the screen, app, page, or mobile camera feed you choose to share, then use Copilot Voice to answer questions and guide you through what you are seeing.
Note
Copilot Vision doesn't use data from your Vision session to train our models. Session data may be temporarily retained for up to 47 hours to support user-submitted feedback and is automatically deleted afterward if no feedback is submitted. A transcript of your conversation with Vision is saved in your chat history so you can revisit it later. Your conversation history isn't used for training or personalization, and you can delete it at any time.
Copilot Vision availability might vary during rollout. Vision is available in all supported regions and languages and can be accessed from Windows, Edge, and the Copilot mobile app on iOS and Android. You must subscribe to Microsoft 365 Personal, Family, or Premium to access Copilot Vision.
How Copilot Vision handles your data
Copilot Vision is only active when you start a Vision session. It handles your data in accordance with Microsoft's Privacy Statement.
When you use Vision, Copilot responses may be logged to help monitor unsafe interactions and outputs. User inputs, images, audio, and shared page content aren't used to train models or personalize your experience. Session data may be temporarily retained for up to 47 hours to support user-submitted feedback and is automatically deleted afterward if no feedback is submitted.
Vision doesn't interact directly with your PC, the web, or your phone on your behalf. It is there to answer questions and guide you, not to take actions for you. Copilot Vision might highlight portions of the screen to help you find relevant information, but it doesn't click, enter text, or scroll on your behalf.
The first time you use Vision in Windows, macOS, Edge, or Copilot mobile, Copilot will ask you to acknowledge a privacy notice.
Vision is available to commercial users signed in with their Entra ID in the Copilot web and mobile apps.
Use Copilot Vision on Windows
Copilot Vision lets you share selected browser windows or apps with Copilot during an active session. Copilot can answer questions about what you choose to share and provide step-by-step guidance.
While a Vision session is active, Copilot can answer questions about the app or browser windows you have shared, even if they are no longer open in front of you. When the Vision session ends, Copilot immediately stops observing the content you shared, and the shared context is no longer available to the active Vision session.
Copilot Vision works with any app on Windows. However, if the content includes harmful or DRM-protected material, Copilot can't analyze it.
To use Copilot Vision on Windows:
- Start a Voice call in Copilot. Vision is off by default.
- To add Vision to the Voice session, select Share screen.
- Select the screen or app you want to share with Copilot. While Copilot Vision is active, you'll see an outline around the shared screen or app. You can share up to two apps at a time.
- Make sure your microphone and audio are enabled. Copilot Vision uses voice input and output to answer questions about the content you're sharing.
- A floating toolbar appears with Vision and Voice controls. Copilot will greet you, and you can begin asking questions about the shared content.
- To mute yourself, select the microphone icon
. This stops Copilot from listening. Select the microphone icon again to resume. - To end the Vision session, select Stop in the floating toolbar. When the session ends, Copilot immediately stops observing the shared content.
Use Copilot Vision in Edge
Copilot Vision works with most sites and pages, except those containing harmful or adult content. When you visit a site that Copilot Vision doesn't support, the Copilot composer at the bottom of the screen will show a grayed-out icon, and Copilot won't discuss the content of the page.
To use Copilot Vision in Edge:
- Sign in to Edge with your personal Microsoft account.
- Select the Copilot icon in the Edge toolbar.
- Select the Voice (waves) button
. - Start a Voice session with Copilot.
- Select Share screen to add Vision to the Voice session.
- Go to any supported website, video, or PDF you want to ask Copilot about. When you're using Vision, a toolbar in the Copilot side pane lets you know your Vision session is active. Vision stays with you as you browse from page to page until you're ready to end the session.
- Ask Copilot questions about the page or content you are viewing.
- To stop the session, select X in the toolbar at the bottom of the Copilot window, close the Copilot side pane, or exit Edge.
Use Copilot Vision on mobile
Note
Mobile Vision features are rolling out soon. Mobile Vision and screen share are being gradually enabled on iOS and Android devices, so availability might vary by user during rollout.
You can access Copilot Vision in the Copilot mobile app on both iOS and Android devices. Copilot Vision answers questions based on what it sees through your mobile device camera or screen.
To use Copilot Vision on mobile:
- In the Copilot mobile app, start a Voice session with Copilot, then select Share screen. This will start your mobile device camera.
- You need to give the Copilot app permission to access your device camera to use this feature.
- Talk to Copilot and ask questions about what you see through the camera or on your screen. Copilot Vision stays on until you end the Vision session.
- To end the Copilot Vision session, select X.