Docs Dashboard

Overview

Bing Visual Search is an image based search service available to millions of users across variety of surfaces. You can use an image as input for search and get rich knowledge and actions. Leveraging our computer vision recognition and knowledge we provide, you can build visual skills that provide new knowledge and actions. Your skills have access to users across Microsoft 1st party products and services built upon Bing Visual Search API.

How does it work?

  1. After the user takes a photo, Bing Visual Search runs it through the image understanding and intent service.
  2. Depending on the detected intent, the service requests different skills.
  3. Then, all skill results are combined and sent to any supported Bing Visual Search clients.

Diagram shows how it works

How to build a skill

  1. Decide when your skill should be invoked. You can configure your skill to be triggered on every image or you can customize the trigger based on the visual intent. To customize the trigger you can build a simple logical expression to combine multiple intents.
  2. Choose additional image metadata that your skill can leverage to be successful.
  3. Create your own web service that will provide users with more knowledge and/or actions. Provide access to your secure endpoint so that we can communicate with your skill.
  4. Test and share your skill with friends and family or other developers. Set up your skill preview and use the test code to try it out in the available platforms.
  5. Publish your skill publicly in the Bing platform. Once your skill is reviewed and approved it will be published for the world to see. Publishing is currently unavailable but will be soon.

Where to find published skills?

In Alpha stage, you can build your skill on the developer portal and test your skill in the Bing iOS app. Support for publishing your skill to be available publicly is coming soon.