Prehension lets you easily implement gestures in your VR games

Prehension AI is simplifying how developers handle animated hand gestures in virtual and mixed reality. Implementing complex hand movements that happen over time is a significant challenge for software engineers because user performance of these gestures varies wildly. Traditional methods struggle to account for this inconsistency, but this new Unity SDK offers a practical path forward.

The development process is straightforward. Developers record their chosen gestures through a Unity scene while wearing a headset. By gathering samples from multiple people, the system builds a library of movement data. This data is then sent to the cloud to train a machine learning model. Once the training completes, the model downloads to the project, allowing it to function offline without a constant connection.

During hands-on tests at AWE, the system showed impressive accuracy. Even when testing gestures not specifically included in the initial training samples, the classifier correctly identified the movements. It also distinguishes between intended gestures and random hand motion, meaning it does not force incorrect classifications on the user. This reliability suggests the underlying technology is well-constructed.

While the software is currently in private beta, it presents a clear value proposition by automating tasks that would otherwise require deep machine learning expertise. The main considerations moving forward remain the future subscription cost and the ability of the company to gain enough market share among independent developers during a difficult period for the hardware industry. If the pricing structure aligns with the needs of smaller studios, this tool could become a standard addition to the developer workflow.