Przejdź do treści


The ultimate generative
3D API toolkit


Generate 3D models, materials, and HDRIs with the Gen3D
toolkit from Shutterstock, powered by NVIDIA Edify.
Now available in beta.

Developed by trusted 3D partners
Built with decades of 3D experience for studios and enterprises.


SSTK+NVIDIA

Generative 3D for artists and studios

Your creative partner for production-ready 3D assets.

HDRI Gen 3D API LP
Models
Create 3D models from text or images in a variety of interchange formats, compatible with all major 3D applications.
HDRI
HDRIs
Illuminate 3D environments day or night, up to 16K resolution, as 32-bit .exr or .hdr outputs for deep exposure control.
Materials
PBR materials
Create realistic or stylized PBR material sets, including 4K diffuse, roughness, and metalness maps.
feature mesh

feature hdri

feature variation

feature uv

div 01
Download our plugin for Blender or our standalone app to get started.

Generative 3D for developers

Easily scale and integrate 3D AI capabilities into any app or experience.

link unified management
Unified management
A one-stop shop for all 3D-related media types with unified and coherent API schemas under one credit system.
link helpful guides and biolerplates
Helpful guides and boilerplate
Learn everything you need with helpful guides and boilerplate for top 3D apps to get started in minutes.
link adaptable to your needs
Adaptable to your needs
Build confidently with flexible plans and an infrastructure backed by NVIDIA.
div 02
Make your first API call and generate content in just two minutes.

Generative 3D for enterprises

Industry-leading capabilities you can trust to lower production costs.

link responsible ai
Responsible AI
The only 3D model generator trained on licensed, ethically sourced data for a 3D solution you can trust.
link cost efficiency and speed
Cost, efficiency, and speed
Reduce 3D production costs with seamless integrations that save you time.
link production at scale
Production at scale
Quickly generate large libraries of varying 3D content to fill virtual worlds.
div 03
Unlock priority generation, fine-tuning, and custom features.

Pricing and credits

Explore credit packs that start at $25 for 2500 credits and scale to your production needs.

3D Workflows

Description

Price (per call)

Text-to-3D Draft

From a text prompt, generates a preview of a 3D model for fast iteration.

10 credits

Text-to-3D Model

From a text prompt, generates a 3D model.

300 credits

Image-to-3D Draft

From an image, generates a preview of a 3D model for fast iteration.

10 credits

Image-to-3D Model

From an image, generates a 3D model.

300 credits

Text-to-Panorama

From a text prompt or image, generates a 4K 8-bit 360 spherical panorama.

100 credits

Panorama-to-HDR

From a provided 360 spherical panorama, generates a 16K 32-bit panorama.

500 credits

Download sample generations

Checkout these sample 3D generations and HDRI panoramas.

download dystopian character
Dystopian explorer
Text-to-3D generation
download desert hdri
Desert canyon pano
Text-to-Panorama and Panorama-to-HDR
download orange chair
Minimalist lounge chair
Text-to-3D generation
download blue rocks
Blue rocks
Text-to-3D generation

What does it mean that this service is in beta?

We are providing users early access to cutting edge capabilities while we gather feedback to drive improvement. We’re actively iterating on features based on input to deliver the best experience. This also means that some features, capabilities and pricing may change as we refine and improve our services.

We will actively communicate any large scale changes and provide a transition period to ensure our users have adequate time to accommodate those changes.


How can I use this API?

Our generative API service can currently be used to generate 3D focused content including 3D models and HDRI panoramas. We expect to grow our offering to include additional media types services in the future.

As this offering is API based, you must make API requests as outlined in our documentation.

To get started, we provide both robust documentation with examples using both cURL and Python.

However, you can also download our plugin for the 3D application Blender, or download our standalone app to start generating in a matter of minutes. We provide guides on how to get started with our integrations in our documentation. We plan on offering more plugins and integrations in other top 3D software applications in the near future.

To generate you will need to purchase a credit pack as well as generate an API token. We provide a step-by-step guide in our Quick Start guides available within the developer platform.


What are the capabilities of the generative API?

We currently support generating 3D models and HDRI panoramas in a variety of formats using both text prompts and images. We call these our ‘workflows’.

We will share more about bringing more workflows across more media types to the service later this year.


Do you provide plugins?

We currently provide a Blender plugin as well as a standalone app. You can learn more about using those in our documentation.


What is preview mode?

Preview mode is an option for text-to-3D and image-to-3D workflows that allow for the quick and cheaper generation of a preview image before you commit to a 3D model generation.

We provide options to receive your previews in images, or gif or .mp4 animated turntables.

You can learn more about preview mode in our documentation.


What file formats are provided for generations?

For 3D generation we provide options for common interchange formats including GLB, USDZ, OBJ and PLY. This enables the 3D models you generate to easily be opened in nearly any 3D application. We also support specifying receiving multiple formats for each generation.

For panoramas, we provide .jpg and .png output formats for our 4K panoramas and .exr and .hdr for our 16K resolution 32-bit panoramas.


Can I generate based on an image?

We support generating both 3D models and panoramas from a reference image.

For 3D model generation, the user provided image serves as a heavily weighted reference to steer the generation. The model may deviate from the image considerably.

For panorama generation, the user-provided image serves as the center of the generated full panorama, with the AI model outpainting to generate a full spherical 360.


What protections or guardrails are in place for this service?

Shutterstock is proud of our ethical approach to AI. You can learn about our overall TRUST framework that guides our approach here.

When it comes to safeguards, we take a multi-layered approach to prevent the creation of offensive, misleading or otherwise inappropriate content. We also have systems in place to help prevent the creation of brand infringing content. These safeguards, while robust, are imperfect but we strive to improve them on a constant basis.

If your request inadvertently triggers our safety or brand detection systems, you are provided an error message and your credits are not consumed.

We take safety and responsible AI usage seriously. Abusive, intentional misuse, or attempting to bypass the safety protections of our services will result in your service being suspended.


What data was used to train the Shutterstock 3D generative model?

The model that powers our 3D generative services was trained on hundreds of millions of licensed images from Shutterstock as well as over 1 million licensed 3D models from TurboSquid.

We add new content over time to continuously improve the quality of our models.


My company is an existing Shutterstock customer, can we get access?

For enterprises we are prioritizing existing customers. To get access, please contact your Shutterstock representative or contact aigenapis@shutterstock.com.

W serwisie Shutterstock.com dostępnych jest ponad 475 000 000 materiałów (stan na 30 listopada 2023 roku).

© 2003-2024 Shutterstock, Inc.