Save a deliberate camera view, combine it with a mood brief, and generate an image you can revisit in Docs. You can complete the camera preparation without generating an image; the render step uses the space's AI credits.
You must be the scene author to render, with space-owner authority or Use AI credits permission in its area. A configured Agent and a readable, nonempty credit balance are also required.
Preserve The View
Save the scene so that the camera becomes part of the document you can return to.
Prepare The Brief
Open Moods and create a named direction. Add a short description of lighting, materials and atmosphere, then choose relevant pictures or articles. Review the linked scene Article too, because its text accompanies the render.
A focused prompt might be Use pale stone and warm evening light. Keep the central path clear and preserve the bench positions. Describe the result you want, including the parts of the layout that matter.
Generate And Review
If the connection fails, use Refresh to check the existing request before starting anything new. A new submission is a separate generation.
Continue The Design
Continue from an eligible completed response to request a focused change. Start a new conversation when you want a separate direction; the saved gallery images remain available.
The generated images are ordinary Docs files, stored in a folder beside the scene. The first output folder uses the scene's title. Editing an image direction does not update the 3D objects, so return to Stage and make any accepted layout changes there.