Status: Soon. Public text-to-image access is pending. When enabled, choose from the models and settings shown in your account. Image creation from a visual reference remains Soon.
Prepare a clear image brief
Start with the subject and its setting, then add framing, light and one or two details that matter. Example: A blue ceramic mug on a walnut desk, three-quarter product view, soft window light from the left, clean background, no lettering. See Prompt writing for revisions.
Use the creation controls
- When enabled, open /studio/image or choose Create → Image inside the editor. Enter your brief in Describe your image (up to 2,000 characters).
- Open Model & output. Choose a model offered in your account and set the options shown for it, such as aspect ratio. The initial V1 text-to-image path uses GPT Image 1; other models depend on account availability.
- Optionally open Help me plan, choose Suggest a plan, edit the composition and use Use this plan. Planning alone does not start image generation.
- Read the current quote and balance before pressing the generation button. If generation is disabled or no model is offered, the account is not enabled for that path; do not assume a hidden fallback.
- Review the result. Choose Download for a file, Edit creation from the standalone page, or Add to this editor when creating inside a project. In the editor, place and time it like another image layer.
Improve a result deliberately
Change one instruction at a time, such as camera angle or lighting, then request a new quote. Regenerate starts a new generation request; it is not a free edit of the existing result.
Can I upload a reference image to guide generation?
The creation UI labels Visual references Coming soon. Use a text prompt or import an existing image into the manual editor instead.
Why is no model listed?
Sign in, check whether image generation is enabled, and refresh if the catalogue could not load. Public activation is still Soon.