Google LLC said today it has added a new “multisearch” capability to Google Lens, enabling users to search for answers using a combination of images and text.

Google Lens, the company’s image recognition technology, allows users to point the phone camera or snap a photo and search Google for answers relating to that image. With the new feature, it becomes possible to take a photo of a shirt, for example, then type to ask Google to find a similar one but with “pinstripes,” for example.

Google, which previewed the capability in September, said multisearch is now available in beta test mode in the U.S., though it only works with English-language users for now.

To use multisearch, users open the Google Search app, then tap on the Lens camera icon to take a photo or choose one from an album. Then, users have the option to swipe up and select “+ add to your search” and enter some text.

Google offered a few examples of what the new feature can do. Users could ask questions about an object that’s right in front of them, for example. They could also refine search results by snapping a photo of an item and asking for a specific color, brand or other visual attributes.

The company said multisearch works best with shopping inquiries but has additional use cases. For instance, if people buy or receive a plant and aren’t sure what it is, they could take a photo and add the text “care instructions” to find out how to take care of it.

The feature could certainly be useful. Suppose a shopper finds a dress she really likes but the store doesn’t have it in the color they want. She could simply snap an image of it and ask Google where to find a “red” version of that dress. Google said multisearch can also do things such as finding a match for furniture. So if someone takes a photo of a dining set and adds the text “coffee table,” Google can find a table that matches it.

Image: Google