The wrnch AI engine helps developers enhance their applications by enabling computers to see human motion, shape, and intent.
Get startedwrnchPC s an ultra-low-latency computer vision engine optimized for the creation of rich interactive experiences. wrnchPC allows you to capture human motion from as many people as you want without the need to wear cumbersome sensors. Our frictionless AI engine is camera agnostic, AI hardware agnostic, built to integrate with all applications easily. We use reliable deep learning models to provide robust human motion capture for all activities and in any environment.
Windows, Linux Ubuntu
versions 16.04 & 18.04
GPU and CPU builds
MacOS iOS
12 and 13 for CPU builds
All NVIDIA GPUs
supporting TensorRT
Intel CPUs supporting OpenVino
C / C++
Python 2.7 / 3.5 / 3.6 / 3.7 / 3.8
UE4 plug-in
Unity plug-in (2018.4 LTS)
TouchDesigner node
wrnch helped Volvox beam visitors to the moon, at the World Trade Center Transportation Hub, aka the Oculus in NYC. Volvox built an interactive experience to mark the 50th year anniversary of Apollo 11’s landing. With wrnch’s technology, Volvox created an entire pipeline for markerless tracking and game engine, offering visitors the opportunity to “walk on the moon”.
LEARN MOREwrnchCloud enables you to transform human pixel data into actionable human-centric insight. This on-demand solution can extract human pose and motion data from any video input. WrnchCloud is a cost-effective solution, scalable, and future proof, for all your human motion data needs
N/A
Sits on top of GCP
Intel CPUs: Skylake, Cascade Lake
NVIDIA GPUs: K80, P4, T4, P100, V100
Other systems upon request
RESTful API
Python 3.5 / 3.6 / 3.7 / 3.8
The wrnchAI platform is at the heart of Intel’s 3DAT: a 3D Athlete Tracking computer vision system to enhance the viewing experience of athletic fans and featured at the 2020 Tokyo Olympics Games. wrnchAI runs lightning fast on Intel® Xeon® processors, to deliver near real-time biometric insight and overlay visualization for sprinting events’ broadcasts.
LEARN MORE
wrnchMobile allows you to transform mobile devices into a real-time motion capture system. Create interactive applications in real or augmented reality. Bring new experiences to life.
All iPhones supporting coreML
Android coming soon write [email protected] for more
C / C++
Unity plug-in
UE4 plug-in
GiantBear, a Hong Kong-based software developer is helping elderly patients suffering from chronic knee pain get better. They developed a human-centric mobile application which uses wrnchAI to capture body movement. Their application can assess biometric information, monitor treatment progression and recommend a personalized exercise plan tailored to each patient’s specific needs.
LEARN MOREwrnchEmbeded gives automated devices, including robots and personal assistants, the power of sight so that they can see and interact with humans in real-time. wrnchEmbeded helps devices anticipate human behavior and respond best in all situations.
Ubuntu Versions 16.04 & 18.04
GPU and CPU builds.
NVIDIA JetPack 4.2 for Jetson™
Nano, TX2 and Xavier
Other systems upon request
C / C++
Python 2.7 / 3.6
MRMC develops assistive robotics and automated technologies for live broadcasts, in the studio and at the game. wrnchAI enables their recording systems to track body and face movement, allowing their cameras to automatically stabilize and deliver smooth broadcasts.
LEARN MORE
