Skip to main content


Megh VAS SDK allows developers to optimize video (and data) analytics pipelines and integrate highly customizable AI into applications and products. It speeds up time-to-market by bringing custom video analytics solutions from proof-of-concept to stable production deployments. Finally, it is a central part of Megh's Open Analytics platform, which allows for building custom, high-performance intelligent video analytics solutions.


VAS is powered by the Nimble application framework, allowing customization across the entire analytics pipeline for use case-specific accuracy. Engineered to seamlessly leverage hardware acceleration for CPU, GPU, and FPGA platforms, the framework makes VAS a truly high-performance, real-time analytics solution.

VAS SDK Pipeline


VAS SDK exposes simple, well-defined and documented APIs and interfaces to achieve these key features:

  • Easy integration of custom algorithms and analytics pipeline stages into the open-source VAS Pipeline Elements and Analytics Library
  • Seamless integration of custom deep learning models into the open-source VAS Inference Library via a flexible deep learning backend infrastructure
  • Quick and easy deployment with the newly integrated custom models and pipeline stages enabled by RESTful APIs
  • Support for standalone inference-as-a-service deployments