Shutterstock launches new computer vision API product to help developers and businesses easily enrich visual search in their product

As Shutterstock continues to invest in AI-powered search innovations, we want to keep our platform open for developers and businesses to build feature-rich applications with Shutterstock capabilities.

To that end, we’re launching a new computer vision API product that will deliver Shutterstock computer vision capabilities via our public API.

Learn more about the launch and see some examples of what you can build.

Subscribe to receive the latest API product announcements.


What’s New

Image by Ozz Design

All Shutterstock’s AI-based visual search capabilities are powered by a convolutional neural network trained with over 290 million high-quality images and video assets. 

Our machine learning technology analyzes each image and video on a pixel-by-pixel basis. This allows us to deliver precise results that correlate to the original image or video in color, composition, content, and effect. 

The new computer vision API product will allow developers to access:

  • Reverse image search
    This allows your end-users to begin their search with an image and receive similar images in return. We built Reveal, a Chrome extension, using our reverse image search, which you can try for yourself.
  • Reverse video search
    Similar to the above, end-users can input an image and receive video clips that look like the image they submitted.

Our documentation, API reference, and Javascript SDK are updated to support the new computer vision product. 


What You Can Build

Image by Shinonome Production

“I’ll know it when I see it” is a common attitude when it comes to finding the perfect image or video. After all, text can’t always perfectly capture the feeling that an image or video evokes. 

Therefore, supporting our customers means understanding and delivering a solution that fits how they want to work. That’s why we’ve developed AI-powered search capabilities that release Shutterstock customers from the constraint of the text-based search box.

With this computer vision product, you can provide your end-users with the same freedom to search visually and with greater precision. The possible applications are broad, but here are some ways you might use this endpoint to enhance your product. 

  • AI-generated video content
    Let’s say you created an application that automatically creates video ads based on a specified image or video from an end-user. Perhaps, the end-user needs a localized video ad that shows scenes from a particular city or region. 

    Using Shutterstock’s computer vision technology, you can enrich your application’s video creation capabilities with high-quality images and videos from Shutterstock. 
  • Automatic image or video curation
    If you have a tool that helps end-users create ad campaigns, you can help design and marketing teams go to market faster by generating content collections based on an image or video they’ve already approved.
  • Train visual models
    Our computer vision features can also be helpful for developers who want to train machine learning models to recognize specific attributes in images and videos.

    For instance, if you want your model to recognize dogs, you can use reverse image or video search to quickly find more images and videos with dogs. 

How to Access the Computer Vision API Endpoint

Image by Jackie Niam

Visit the developer portal to create an app and contact us to get access to the computer vision product. 

Once we add computer vision access to your API key, you’ll be able to test and integrate it into your application. The approval process is brief and is implemented for security purposes. 


What’s Next

Our goal is to increase accessibility to the Shutterstock platform so developers and businesses can innovate with Shutterstock content and technology. 

In the future, we plan to release additional computer vision features. If you have specific suggestions, please let us know

We’re excited to see what you’ll create with Shutterstock.