Our librairies and APIs provide developers with an easy way to implement semantic detection on a camera videofeed, in real-time.

For example, by using our “faceFilter” API, a web developer can build a camera-based webapp featuring AR lenses and filters just as Snap or Facebook. All without any knowledge of 3D programming, machine learning / deep-learning or otherwise!

We design and train artificial neural networks, fully programmed in javascript/webGL from our unique framework, to solve very specific problems in the constrained real-life environment. For instance, we trained neural networks to understand human emotions from the webcam of modern smartphones. Launching the API soon, it could be used to build talking avatars (an animoji-like feature) in the browser, for video-conferences or just messaging.