Geometric deep-learning technology company Physna has launched the search engine Thangs using deep-learning algorithms to index 3D models based on geometry, rather than text and images. It allows anyone to search for three-dimensional objects similar to Google’s technology, where users can search by 2D images.

“We live in a 3D world, but digital technology is two-dimensional,” said Paul Powers, CEO of Physna, the company he founded in 2016.

Thangs users upload parts and receive suggestions for where that part may be used and what commercially available components are compatible. Users can search based on the object’s physical properties, measurements and features.

Today, this version is free. An enterprise version has additional features that requires a paid subscription service.

Developers are working to implement an advertising model where brands can participate in a bidding auction, similar to Google Ads, to advertise their parts. Powers said to expect the advertising model in 2021.

The search feature will soon allow consumers to upload a 2D picture to find a part. The technology will show the user where to buy the part and how to install it. In the future, it also will predict what type of 3D printer the user needs to print a model.

“If you manufacture products for use during coronavirus and your vendor shuts down, this search engine can show you different vendors from which to procure your parts in place of the other,” he said.  

Powers works with Physna CTO Dennis DeMeyere, who spend nearly eight years at Google, most recently in the Office of the CTO and technology director for Google Cloud at Alphabet.

Physna, originally designed for designers and engineers, is being used by other companies and agencies. In addition to working with the U.S. Department of Defense, consumer product goods companies use the search engine to make predictions about models and packaging.

The company’s technology is a blend of Google and Microsoft’s GitHub for three-dimensional objects.

In addition to the capabilities of popular 3D-printing and CAD tools like Thingiverse or Yeggi, which allow users to upload and search for models of things to print, Thangs also provides a graphical, collaborative history of each model, similar to the edit history on a Google Doc or version controlling on GitHub.

The free tool allows users to set up a profile, comment on and collaborate on any public model, and store the models.

Physna takes three-dimensional data and finds any 3D relations such as similar models and components that fit into other parts. Think physical DNA and trinary code, Powers says.

The technology codifies of any real-world object and translates it. A series of algorithms extracts all information about a model and reassembles it in code the search engine can read.