Stable Point Aware 3D
Stable Point Aware 3D (SPAR3D) can make real-time edits and create the complete structure of a 3D object from a single image in a few seconds. SPAR3D combines the strengths of point-cloud diffusion (probabilistic) and mesh regression (deterministic) to have improved details on the unseen back regions in the input image.
Compared to our previous model Stable Fast 3D, this new one allows editing of backside information using the point cloud representation and also leverages a larger Diffusion model to generally improve the depth and backside predictions.
Read more about the model capabilities here.
This API is currently in preview. Please don’t hesitate to contact us with any questions.
How to use
Please invoke this endpoint with a POST request.
The headers of the request must include an API key in the authorization field. The body of the request must be
multipart/form-data.
The body of the request should include:
image
The body may optionally include:
texture_resolutionforeground_ratioremeshtarget_typetarget_countguidance_scaleseed
Note: for more details about these parameters please see the request schema below.
Output
The output is a binary blob that includes a glTF asset, including JSON, buffers, and images. See the GLB File Format Specification for more details.
Credits
Flat rate of 4 credits per successful generation. You will not be charged for failed generations.
Authorizations
Your Stability API key, sent as a bearer token in the Authorization header.
Headers
Body
Determines the resolution of the textures used for both the albedo (color) map and the
normal map. The resolution is specified in pixels, and a higher value corresponds to a
higher level of detail in the textures, allowing for more intricate and precise rendering
of surfaces. However, increasing the resolution also results in larger asset sizes, which
may impact loading times and performance. 1024 is a good default value and rarely requires
changing.
512, 1024, 2048 Controls the amount of padding around the object to be processed within the frame. This
ratio determines the relative size of the object compared to the total frame size. A
higher ratio means less padding and a larger object, while a lower ratio increases the
padding, effectively reducing the object’s size within the frame. This can be useful when
a long and narrow object, such as a car or bus, is viewed from the front (the narrow
side). Here, lowering the foreground ratio might help prevent the generated 3D assets from
appearing squished or distorted. The default value of 1.3 is good for most objects.
1 <= x <= 2Controls the remeshing algorithm used to generate the 3D model. The remeshing algorithm determines how the 3D model is constructed from the input image. The default value of "none" means that the model is generated without remeshing, which is suitable for most use cases. The "triangle" option generates a model with triangular faces, while the "quad" option generates a model with quadrilateral faces. The "quad" option is useful when the 3D model will be used in DCC tools such as Maya or Blender.
none, triangle, quad Response
Generation was successful.
The bytes of the generated 3D model.