The complete platform for computer vision
  • Capture 2D & 3D data from edge devices
  • Analyze with machine learning
  • Label for training with our integrated tools
  • Deploy custom models
  • Collaborate across your team
We power perception for the world’s biggest companies and startups alike
  • Ecommerce

    Sizing & Customization

  • Manufacturing

    Digital Twin

  • AR/VR

    Content Creation

  • Retail

    In-store experiences

  • Media

    Visualize 2D and 3D data

Get going faster with Templates for your use case →
Our tools
The edge and cloud tools you need
Capture data at the edge
Collaborate in the cloud
Analyze it anywhere


Capture 2D images and 3D scans with our mobile SLAM pipeline.

Use ML to analyze your scans including semantic segmentation, landmarking, and classification.

Many more tools for data management, networking, file IO, and classical algorithms for processing scan data.


Our web platform gives your team a workspace to store, organize, label & collaborate on data and models.

View the results of edge processing, design, validate, and run tests in the browser.

Our JS SDK makes it easy to drive from your local computer or backend.

Learn more about our product →

Including our open source projects

Our Approach
Extending Perception

The proliferation of new sensors in edge devices - iPhone, Kinect, Galaxy Note, Pixel, and countless others - is changing the way our devices see the world. To date, the data is difficult to gather and make sense of. We’re building the vision and learning tools to help teams of all sizes and skill levels see and understand the world around them.

There are already more than 100M depth sensors in pockets around the world. New sensors mean new possibilities. What will you build?

By Developers, for Developers

  • Clean, sensible APIs
  • Easy customization for your use case
  • Seamless integration
  • Use as tools or end-to-end

Always Improving

  • New features every month
  • Growing list of edge device compatibility
  • Expanding set of pre-trained models
Ready to get