An AI render turns a captured Stage view into an image guided by your brief. Use it to explore materials, atmosphere or presentation while keeping the editable 3D scene as the basis for your next decision.

Build A Brief You Can Reuse

Open Moods from the Render menu to create a named visual direction. A mood can contain a description, Docs pictures and article references. For example, Evening courtyard might describe warm light and pale stone, with a small set of pictures that clarify those choices.

A scene can hold up to 32 moods, each with up to 12 pictures and 12 articles. Picture upload accepts up to 12 files in a batch. Removing a picture or article from a mood leaves its source document intact.

Mood boards save with the scene. The currently selected mood is a local choice, allowing you to prepare different directions without changing the content of the other moods.

Decide What To Send

The render conversation combines the chosen camera capture, your prompt, the scene's linked Article and the selected mood's description, pictures and current article text. Review these together: an old sentence in an article can undermine a more recent instruction in the prompt.

In the Chat panel, choose a Camera, Mood, Agent and Output format. Current viewport uses your current view; a named camera uses its saved frame. WebP and JPEG are the supported output formats. The available Agents depend on the configured models shown in the app.

Rendering makes images. It does not rebuild the 3D scene from the result.

Understand Access And Credits

Only the scene author can start a render. They also need authority to spend the space's AI credits: space-owner authority or the area's Use AI credits permission. Scene access alone is not enough.

Stage shows the available allowance as a whole percentage. Credit usage is tracked during generation and adjusted to the final recorded cost, so there is no single fixed credit price for every render. Rendering stops if the shared pool is exhausted. A balance that cannot be checked appears as unavailable and prevents a new request.

Follow The Result

Progress distinguishes queued work, generation and saving from a completed result. A conversation can continue from an eligible completed response, so you can ask for a targeted change instead of starting the discussion over.

Refresh after a connection problem to check the existing request. Repeating the prompt as a new request is a separate generation. Stop requests cancellation; keep watching the status if cancellation cannot yet be confirmed.

Saved images appear in the render gallery and as ordinary Docs files. New output is placed in a folder beside the scene, initially named after the Stage document. The gallery also includes captured scene views. Starting a new conversation does not remove these files.

Common Questions

Can other viewers read my render conversation?

Render conversation history is scoped to your account and that scene. Saved image files have their own normal Docs access, so their audience can differ from the conversation's audience.

Can I still see old renders when no Agent is available?

Yes. The gallery can show previously saved images even when a model is unavailable.

Does Stop guarantee no credits were used?

No. Work may already have run before cancellation. Usage is accounted for during generation and settled from recorded cost.

See save a camera and render an image for the connected workflow.