Start your first 3D creation with a reference image
Choose a clear subject, describe what matters, then inspect the result before your next iteration.
A clear reference is the starting point for image-to-3D creation. Use a jewelry photo, a sketch or a design image; you do not need to learn every generation parameter first.
Choose a clear subject
Keep the whole subject visible, without hands, text or other objects covering it. A simple background and a clear silhouette help communicate the form. Strong reflections and small openings may need closer inspection in the result.
Describe what you want to preserve
After uploading, try: “Turn this image into 3D and preserve the twist in the middle.” The assistant uses the image and your description to prepare a plan. You can also explicitly select image-to-3D and adjust the supported parameters.
Review the reference, parameters and estimated credits before confirming. Conversation and generation are charged separately; confirming a generation plan starts its task. A written description of a model is not a completed model.
Inspect the preview
Rotate the result to view its front, sides and back, then zoom in to check the silhouette and details. A single image does not show every surface, so generated geometry can differ from the real object.
To try another version, reuse the reference and parameters from a result and describe the next change. This creates a new generation task; it does not directly edit the existing mesh.
Continue with the model file
You can revisit delivered results in your works and download the model file. Before prototyping or production, check dimensions, wall thickness, structure and manufacturability in suitable tools. Generation does not perform those checks for you.
If you upload several independent images, explain whether each needs its own model. Multiple views of the same object are a different request and should not be treated as separate objects.
For the steps and current supported scope, see the image-to-3D guide, or start creating.