V0: Annotations Mode for Visual Feedback
V0V0 introduced Annotations Mode, a visual feedback layer that lets users click directly on preview elements to drop numbered comment pins, then send the entire batch to the agent at once. Combined with the ability to attach images to queued prompts while another chat is still running, the June 19 release marks a meaningful step toward more visual and asynchronous interaction with V0.
Sources & Mentions
4 external resources covering this update
Overview
V0 has rolled out Annotations Mode, a new visual feedback layer that lets you communicate changes by pointing directly at the preview instead of describing them in text. The June 19 release shifts a meaningful part of the workflow away from the chat box and onto the rendered UI itself.
Click to Annotate
With Annotations Mode active, you can click directly on any element in the preview to drop a numbered comment pin. Each pin anchors a note to a specific part of the interface, so feedback like "tighten this spacing" or "change this label" is tied to exactly the element you mean. The numbering keeps multiple pins organized and easy to reference.
Batch Send to the Agent
Rather than sending each comment one at a time, you can collect a full set of annotations and send the entire batch to the agent in a single pass. This lets you do a complete review of a screen, mark up everything you want changed, and then hand the agent one consolidated set of instructions to act on.
Image Attachments on Queued Prompts
The same release adds the ability to attach images to queued prompts while another chat is still running. You no longer have to wait for the current generation to finish before lining up your next request with visual references, which keeps the back-and-forth moving.
Why It Matters
Together, these changes push V0 toward a more visual and asynchronous way of working. Pointing at the UI is often faster and less ambiguous than writing prose, and queuing image-backed prompts means you can keep planning while the agent works.