Abstract: Systems and methods for video processing generates 3D streaming data via an API may be provided. A 3D streaming session may be delivered at a client device with low latency and high throughput capacity. Upon verification of the client device, a server may be selected to support an encapsulated 3D streaming engine to initiate the session. A library file may be injected into an executable file of the encapsulated 3D streaming engine to generate a digital representation of an interactive 3D environment. The interactive 3D environment may be rendered, encoded, and streamed to the client device via a GPU. The 3D streaming session may be adjusted via the client device, and sustained until terminated at the client device. A number of virtual compute instances of encapsulated 3D streaming engines provided by the server may be dynamically adjusted based on metrics and machine learning predictions of connected client devices.