SDK Quickstart Installation

The CineStream SDK provides direct C++ and Python bindings to Pixar OpenUSD and the NETUNISYS GPU cluster orchestration layer.

# Install via Python PIP

$ pip install cinestream-sdk omniverse-openusd

# Install Node.js Core Client

$ npm install @cinestream/core

Authoring Scenes in Python

Initialize a connected stage, bind Audio2Face rigs, and dispatch RTX raytracing with less than 15 lines of code:

import cinestream as cs
from cinestream.audio import Audio2Face

# 1. Authenticate with NETUNISYS Cloud Cluster
client = cs.Client(api_key="cs_live_99x8f0a21b")

# 2. Instantiate or connect to OpenUSD stage
stage = client.stages.create(
    name="Hero_Cyber_Scene",
    template="cinematic_metropolis_v2"
)

# 3. Bind digital actor with Audio2Face inference
actor = stage.spawn_actor(
    rig_uri="omniverse://fabric.cinestream.in/Actors/hero.usd",
    audio_track="./audio/hero_dialogue.wav",
    engine=Audio2Face.OMNI_TENSOR_V2
)

# 4. Stream 4K RTX Raytraced Viewport
stream = stage.render_stream(resolution="3840x2160", fps=120)
print(f"Stream Active: {stream.websocket_uri}")

OpenUSD Stage Schema Specification

CineStream extends standard Pixar OpenUSD with custom Schemas for real-time neural weights and physics boundary conditions:

CineActorSkelAPI

Encapsulates ARKit 52 blendshape target mappings, phoneme confidence thresholds, and micro-saccade eye rates.

CinePINNsVolumeAPI

Specifies fluid viscosity, boundary aerodynamic turbulence, and soft-tissue elastic modulus for continuous physics.