Cursor Design Mode: Multi-Select and Voice Input
Cursor 3.7 expands Design Mode in the in-app browser with two significant improvements. Developers can now select multiple UI elements simultaneously, giving the agent full visibility into their code, layout, and visual relationships to make coordinated cross-component edits. A persistent voice input overlay is also available throughout agent runs, allowing users to narrate changes or queue the next instruction by voice without waiting for the current operation to complete.
Featured Video
A video we selected to help illustrate this changelog
Sources & Mentions
2 external resources covering this update
Multi-Element Selection
One of the most common friction points in AI-assisted UI editing is the difficulty of explaining which element to target. Cursor's June 5, 2026 update to Design Mode directly addresses this by introducing multi-select in the Cursor browser. Developers can now click two or more elements simultaneously, and Cursor surfaces the selected elements' code, the surrounding layout, and the visual relationships between them all at once.
This richer spatial context allows agents to make coordinated changes across components in a single pass. Practical use cases include matching the styling of one element to another, removing visually duplicated content, or adjusting an entire group of components in a way that respects how they relate to each other on the page.
Persistent Voice Input
Design Mode already supported voice narration as a way to describe UI changes without typing. Cursor's 3.7 update makes voice more useful by keeping the microphone available while an agent is actively processing a previous request.
Previously, users had to wait for an agent to finish before they could describe the next change. Now the microphone stays on during agent execution, so the next instruction can be queued by voice mid-run. This reduces idle time in iterative UI editing sessions, particularly when each change prompts the need for a follow-up tweak.
Canvas Support
Design Mode is no longer limited to the Cursor browser. The same multi-select and voice capabilities are now available in canvases, extending the visual editing workflow to Cursor's canvas surfaces where agent-built interfaces are previewed and refined. Developers can select and annotate UI elements directly in a canvas to guide edits, just as they would in the browser.
Bug Fixes
A number of stability improvements shipped alongside these features, tightening the accuracy of element targeting and refining the performance of visual interactions in the browser overlay.