The Most Efficient AI Video Analytics

Reduce Cost
by 99%

For customers to optimize profit and to consume less energy. And bring more cost-effective services to the mass market.

Improve Performance
by 100X

For customers to improve their performance. And even create new products that are not technically feasible before.

How we help

Improve cost efficiency at edge by 200X

Improve throughput at work station by 100X

Improve cost efficiency in cloud by 200X

Speed up visual search through video footage by 100X

Improve storage efficiency by 1000X

How are we different?

Highest performance at lowest cost

Boost performance by up to 100X, reduce operating cost by 99% and improve energy efficiency by 100X, no matter where, whether in cloud or at edge.

Best optimized for legacy systems

Transform ordinary cameras, aging servers and low-bandwidth networks into state-of-the-art video analytics systems, possibly with zero hardware change.

Complete freedom to choose any model

Take full advantage of state-of-the-art open-source models or integrate smoothly your own proprietary or third-party ones without exposing any detail.

100X Performance, 100X Efficiency, 1/100th Cost

Improve performance by 100x throughout ubiquitous computing, whether in cloud, on workstation, at gateway or on device
Reduce hardware expenses by 99% on computing as well as networking, while retaining the same functionality
Improve energy efficiency by 100X and reduce power consumption by 99%, beyond fundamental limit

State-of-the-art open-source model recognizes thousands of types of objects with better than human accuracy

Deep Neural Network based Model recognize thousands of types of objects at better-than-human accuracy
Yet ready-to-use without the need for any tuning for specific application
Open-source model is free to use, free to modify to suit your own needs, free from vendor lock-in
Open-source model conforms to open standard, improve interoperability
Global open-source community provides detailed documents and comprehensive support
With millions of users examining source-code, open-source model has fewer bugs, faster fixes and better security

Cutting-edge video analytics platform readily integrable into your existing software stack and hardware infrastructure, also work with your proprietary model

Ready to apply your proprietary model of limited input resolution to any video stream of unlimited resolution natively without any downsampling
Plug-and-play, smoothly integrable into your existing software stack, no matter where, whether in cloud, on workstation, at gateway or at edge devices
Effortlessly scaling to thousands of legacy cameras and aging servers without the need for additional bandwidth
Hardware agnostic and device independent platform works on smart cameras, edge gateways, network video recorders, on-premise servers. monitoring workstations and cloud computing services

Improve productivity astronomically by advanced video analytics with capabilities approaching human operators and beyond

Visual Search scans through days of video in seconds to locate a specific object (person, vehicle, face, cat and others) of specific description (color, size, speed and so on)
Video Summary removes uneventful scenes while retaining relevant ones, enabling security professionals to review hours of footage in minutes
Real-time Alerts relieve human operators from continual monitoring uneventful scenes, direct their focus to suspicious activity and increase their situation awareness, enables one human operator to monitor hundreds of video streams
Region of Attention directs human operators to micro regions of interest across vast horizons of camera views as they expand to tens of megapixels and 360 degrees
24/7 active monitoring by AI without the need for human operator, and thus without distraction, fatigue, training, or needing to be paid
Truly preventive alarms deter crimes from happening, enable security personnel to proactively respond to suspicious activity

Enable unlimited cameras, highest resolution, highest frame rate, highest encoding quality, all without compromise to limited bandwidth and storage

With low cost (< $100) and low power (<20W) AI-powered video analytics at the edge, only frames of interest and regions of interest need to be transmitted through network to minimize consumption of bandwidth and storage, existing infrastructure with limited bandwidth and storage may be utilized without costly upgrades. Meanwhile, take full advantage of:

4K 8 megapixel and even 8K 32 megapixel resolution providing 4 to 16 times the detail as a 1080p HD camera for clearer images, wider coverage and more accurate analytics
Higher frame rate to capture more actions and reduce motion blur, amplifying the benefits of 4K video
Higher quality encoding to reduce any loss in details for more realistic recording
Prioritize and optimize coverage with as many pixels as you need and as many cameras as you need

Enable AI at the edge for less than $100, delivering great advantages over AI in the cloud

Eliminate latency to transfer data to cloud, enable real-time response
Filter relevant Information, reduce communication overhead
Distribute computing to local, ensure reliable operation not interrupted by connectivity issue
Keep sensitive data at local, avoid security and privacy concern of cloud
Aggregates data locally without transmitting to cloud, largely eliminating networking cost

Revitalize and expand legacy infrastructure into modern age of AI

Transform ordinary camera, whether analog or digital, into state-of-the-art video analytics system through $100 gateway at the edge providing real-time analysis by artificial intelligence
Revitalize low-bandwidth network to support more cameras, higher resolution and faster frame rate through $100 gateway at the edge extracting and transmitting only events of interest
Breathe new life into aging servers and workstations through empowering them with the ability to perform cutting edge AI-powered video analytics in most efficient way