Generated banner for Omniverse Realtime Viewer
Generated logo for Omniverse Realtime Viewer

Omniverse Realtime Viewer

NVIDIAOfficialCreated by NVIDIA Omniverse

  • Skill
  • Tool integration
  • Software development
  • Apache-2.0
  • Medium risk

From the author

Use as the top-level router for Omniverse Realtime Viewer USD apps, XR delivery, SPG GPU processing, stylized rendering, and focused viewer reference documents.
— the description by NVIDIA Omniverse

AutoGineer summary

Written by AutoGineer from the skill’s own files, not by the author

The Omniverse Realtime Viewer skill acts as a central routing and architectural framework for developing USD-based 3D viewer applications. It provides a structured approach to composing rendering, scene state, and delivery mechanisms, ensuring that developers can integrate OVRTX rendering, OVStage scene management, and various UI or streaming adapters effectively. The skill emphasizes a modular architecture where the viewer runtime, UI/input adapters, and delivery systems are decoupled, allowing for consistent behavior across local and streamed environments.

What it does

  • Provides a routing architecture for USD viewer apps, XR delivery, and GPU-accelerated rendering.
  • Defines the separation of concerns between viewer runtime, UI adapters, and delivery systems.
  • Offers guidance on integrating OVRTX for high-fidelity 3D rendering.
  • Establishes conventions for camera interaction, scene state management, and input handling.
  • Supports the implementation of specialized features like stylized rendering and CAE/CFD visualization.
  • Provides protocols for agent-to-UI communication and scene control.

When to use it

  • When starting the development of a new USD-based realtime viewer application.
  • When integrating advanced rendering features like stylized visualization or AOV switching.
  • When building XR or spatial editing capabilities into an existing Omniverse application.
  • When you need to standardize the architecture for local and streamed viewer deployments.
  • When implementing chat-driven scene control or agent-based inspection of a 3D scene.

Advantages

  • Promotes a clean, modular architecture that separates rendering logic from UI and transport.
  • Ensures consistent behavior across different delivery methods, such as local windows or WebRTC streaming.
  • Provides clear paths for extending functionality without disrupting the core viewer runtime.
  • Facilitates the integration of complex simulation and visualization data into the viewer.
  • Standardizes conventions for scene state, selection, and camera control.

Usage examples

“How do I set up a basic USD viewer application using the Omniverse Realtime Viewer architecture?”
The agent will guide you to start with the viewer app reference, resolve dependencies, and implement the core runtime owner to manage the OVStage and OVRTX renderer.
“I need to add stylized rendering to my existing viewer. Where should I start?”
The agent will direct you to the stylized visualization references to implement the required USD/Lua/CUDA nodes within your existing runtime owner.
“How can I enable an AI agent to control the viewer scene?”
The agent will point to the A2UI Scene Copilot reference to implement an operation interface that allows the agent to send commands through the existing runtime owner.
“How do I switch between different render outputs like beauty and depth in my viewer?”
The agent will explain how to use the AOV switching reference to map RenderVars and update the RenderProduct's orderedVars through the runtime owner.

Install

  1. Download the skill as a ZIP:

    https://mp-staging.autogineer.com/agent-skills/nvidia/omniverse-realtime-viewer/download
  2. Unzip it into the cross-client skills folder of your project:

    mkdir -p .agents/skills && unzip omniverse-realtime-viewer.zip -d .agents/skills

Documentation by NVIDIA Omniverse

<!-- SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. -->

<!-- SPDX-License-Identifier: Apache-2.0 -->

Omniverse Realtime Viewer

Build responsive USD viewers by composing OVRTX rendering, OVStage scene state, and the delivery/UI libraries appropriate to the app. Use this package for viewer architecture and interaction; use each selected SDK's upstream guidance for installation and exact release-specific APIs.

Start With The App

  1. When delivery is explicit, select its recipe through references/routing.md and follow its current phase. Otherwise start with references/usd-viewer-app/README.md. For an existing app, keep its architecture and read only the references for the requested change. Broad app, conventions and UI catalogs are lookups as needed.

  2. Resolve the selected libraries through references/dependencies/README.md. Preserve the project's manifest unless an upgrade is requested.

  3. Build a working frame and camera interaction in the chosen shell, then add the requested capabilities. Read references/conventions.md for shared input, camera, selection, and protocol conventions.

  4. Use references/viewer-ux-workflow/README.md for broader interface design. Loading, selection feedback, errors, and input responsiveness belong in the first working app, alongside rendering.

  5. Validate the actual delivery path with references/validation.md. A browser screenshot does not validate a native window or a different transport.

Architecture

OwnerResponsibility
Viewer runtimeOwn the live OVStage, publication clock, OVRTX renderer, scene changes, and render stepping.
UI and input adapterDisplay frames and copied app state; send user intent to the runtime owner.
Storage/query/simulation workersPerform their assigned work and return results through the selected integration boundary.
Delivery adapterPresent locally or stream through ovstream; respect frame ownership and backpressure.

Use OVRTX for USD/3D rendering. Browser apps display its WebRTC video; local apps present its output through their native shell. A missing GPU means rendering is unverified, not a reason to substitute a browser 3D renderer.

Keep viewer cameras, settings, effects, and transient edits in app-owned session or runtime state so the source USD stays intact. Serialize scene mutation and render stepping through the runtime owner. Read references/ovstage-ovrtx-integration/README.md for that loop, then the population or data-plane references when implementing scene loading or attribute writes.

Keep expensive loading and metadata work off the UI and render threads. Keep transport details behind a delivery adapter so camera, selection, properties, and loading state can be reused across local and streamed shells.

Add Capabilities As Needed

Use references/routing.md to select camera/picking, scene management, render settings, cloud assets, simulation, editor tools, deployment, or CAE/CFD visualization. Read the selected capability's implementation flow and its linked contract; the full reference catalog is not a prerequisite for a first frame.

For simulation, use the routing reference's Simulation Ecosystem section plus references/realtime-sim-integration/README.md. For UI-driven forces or control, add references/interaction-round-trip/README.md.

For example scenes, use references/usd-sample-data/README.md. When a runtime failure needs diagnosis, use references/troubleshooting/README.md and the selected SDK's matching release guidance before changing the architecture.

Solid Modeling: Upstream First

For usd-brep, exact BRep, solid modeling, kernel failures, CAD topology, or BrepArray questions, first read references/dependencies/README.md → USD BRep. Follow the owning repository's current agent guidance, skills, operation docs, and package API before selecting local integration references. Discover the upstream router rather than assuming skill names or directory layouts remain fixed. Upstream owns modeling semantics, schema rules, tuning, and diagnostics; this package owns OVRTX display, worker handoff, selection mapping, and publication timing.

GPU Processing And Stylized Viewers

For custom GPU processing, chained outputs and lens sampling, read references/spg-processing/README.md. For toon, ink, technical/X-Ray styles, geometry strokes or full wireframe, continue to references/stylized-visualization/implementation.md. Its USD/Lua/CUDA starting node and geometry/depth/composite build sequence work with the selected viewer recipe; use references/stylized-visualization/style-recipes.md for the preset math and references/stylized-visualization/README.md for supporting contracts. Implement the result inside the existing OVRTX/runtime owner and deliver the published processed output. These instructions do not require a sample checkout or a prior stylized-rendering implementation.

For an optional usd-brep geometry source, first read references/dependencies/README.md → USD BRep and the owning repository's current agent guidance, skills, and operation docs; then read references/stylized-visualization/usd-brep-source.md for the display boundary.

XR Viewers

For headsets, immersive viewers, or spatial editors, start with references/xr-viewer-recipe/README.md. Resolve SDK setup and contracts through references/dependencies/ovxr.md; the library owns OpenXR lifecycle and frame submission. For controllers, hands, navigation, or scene tools, add references/xr-spatial-input/README.md and references/ovxr-python-interactions/README.md. For projection-integrated UI, read references/ovxr-composition-ui/README.md and references/ovxr-composition-ui-validation/README.md.

Files

references/headless-shm-cli/README.md

Headless SHM CLI

Triggers

Use this skill for headless CLI, SHM automation, ovusd-shm, viewer automation, scripted interaction, CLI testing, non-interactive viewer, screenshot capture, automated scene validation, scripted pick/select sequences, or programmatic stage tree inspection.

Use this with an existing OVRTX server that exposes the ovstream shared-memory transport. The CLI is a local automation client: it does not render USD, host a UI, or replace the viewer server. It attaches to the named SHM stream, reads frames, sends input and JSON viewer messages, and exits with deterministic stdout/stderr behavior suitable for scripts and CI jobs.

Purpose

Build a command-line client for automation, testing, CI pipelines, and scripted interactions with a running OVRTX viewer server through ovstream SHM. Typical uses include:

  • capture a rendered frame as PNG, JPEG, or raw BGRA/RGBA bytes

  • verify stream health and frame dimensions

  • inspect the USD stage hierarchy and selected prim properties

  • run scripted camera drags and pick/select workflows

  • switch AOVs for validation frames

  • pipe raw frames into video encoders or image-diff tools

When To Use This

Choose the headless SHM CLI when:

  • an automated test needs to drive a local viewer without opening Electron, Tauri, ovui, or a browser

  • CI needs smoke tests against an OVRTX server with --shm enabled

  • screenshots or raw frames must be captured from the same renderer path as the desktop viewer

  • scene validation needs hierarchy, property, AOV, selection, or pick checks

  • scripted interactions should reproduce user workflows such as click, drag, select, inspect, and capture

  • Playwright or another e2e harness needs a stable backend for renderer state

Do not use this as a remote browser streaming client. For WebRTC browser delivery, use streaming-server, streaming-client, streaming-messages, and streaming-lifecycle. For an interactive local desktop UI, use electron-shm-viewer, tauri-local-viewer, or local-viewer.

Architecture

Node.js CLI
  -> generated local SHM client module
  -> native Node addon built with node-gyp
  -> libovstream_shm_client.so
  -> libovstream.so
  -> POSIX shared memory frames and control messages
  -> Python OVRTX server started with --shm

The Python server remains the source of truth for USD, ovrtx.Renderer, stage queries, camera state, picking, selection outlines, render settings, and AOVs. The CLI is only a client:

  • Frames: wait for the newest SHM frame and write PNG/JPEG/raw output.

  • Input: send mouse move/button/wheel events to the server.

  • Viewer state: send JSON event_type/payload requests over the SHM control channel and wait for matching responses.

Reference electron-shm-viewer for the server-side SHM lifecycle and streaming-messages for message envelope names. Reference object-selection, stage-hierarchy, stage-attribute-reads, and aov-switching when adding server handlers that the CLI calls.

For ovstream SHM acquisition and runtime setup beyond this CLI pattern, read references/dependencies. Keep the CLI client code local to the generated app.

Use the unified native client API and frame ownership rules. Copy borrowed frames before encoding or retaining them. Build the Node addon for Node's ABI separately from the Electron addon.

Prerequisites

  • A Python OVRTX server is already running with SHM enabled, for example --shm.

  • The CLI --stream-name matches the server --shm-stream-name.

  • Node.js 18 or newer is available.

  • LD_LIBRARY_PATH includes the directory containing libovstream.so and any dependent ovstream native libraries.

  • A generated local SHM client module exists in the app workspace, for example clients/shm-client/.

  • The native addon has been rebuilt for the active Node/Electron ABI with node-gyp.

Example environment:

export LD_LIBRARY_PATH=/path/to/.venv/lib/python3.10/site-packages/ovstream/lib:$LD_LIBRARY_PATH
export OV_SHM_STREAM_NAME=ovrtx-viewer

Command Reference

Commands should be named after actions and produce script-friendly output. Successful commands write the primary result to stdout unless --output is provided. Diagnostics and errors go to stderr. Non-zero exit codes indicate failure.

frame

Capture one frame from SHM.

ovusd-shm frame --output frame.png
ovusd-shm frame --output frame.jpg --format jpeg
ovusd-shm frame --format raw > frame.bgra

Options:

  • --output <path> writes to a file instead of stdout.

  • --format png|jpeg|raw controls encoding. Infer png or jpeg from --output extension when possible.

Raw output is tightly packed frame data in the native SHM format unless the package explicitly converts it. Include width, height, format, and pitch in info output so raw consumers can decode it correctly.

info

Print stream and latest-frame metadata as JSON.

ovusd-shm info

Include at least stream name, producer liveness, width, height, pixel format, pitch bytes, and sequence number.

tree

Query stage hierarchy rows from the server.

ovusd-shm tree --root /World
ovusd-shm tree --root /World --json

Options:

  • --root <path> selects the root prim path. Default to /World only when the server has no loaded root-prim state to report.

  • --json prints the raw tree response for machine checks. Without --json, print a readable indented tree with prim paths and types.

click

Send a viewport click using fractional viewport coordinates.

ovusd-shm click 0.40 0.45
ovusd-shm click 0.40 0.45 --wait-select

Arguments are x y in normalized viewport space, clamped to 0..1. --wait-select waits for a selection change and prints the selected prim path or null.

pick

Run a raw pick query without changing selection.

ovusd-shm pick 0.40 0.45

Print the picked prim path or null. Use pickRequest/pickResult on the control channel so the server can distinguish raw pick queries from selection clicks.

drag

Send a scripted viewport drag for orbit or pan behavior.

ovusd-shm drag 0.45 0.50 0.65 0.40 --steps 20 --duration 500

Arguments are x1 y1 x2 y2 in normalized viewport space. Options:

  • --steps <n> controls the number of move events.

  • --duration <ms> spreads the drag over the given duration.

Default to left-button drag for orbit unless the CLI adds an explicit button or mode option that maps to the server's camera controls.

select

Select a prim by absolute USD path.

ovusd-shm select /World/Cube

Send selectPrimsRequest {paths:[path]} and wait for stageSelectionChanged. Print the confirmed path or resulting selected paths.

props

Print properties for a prim.

ovusd-shm props /World/Cube
ovusd-shm props /World/Cube --json

Without --json, print tab-separated name, type, and value rows. With --json, print the server response as structured JSON. Cap large payloads on the server and preserve any truncated flag.

aov

List or switch render AOVs.

ovusd-shm aov --list
ovusd-shm aov --set LdrColor

Options:

  • --list prints available AOV names, one per line.

  • --set <name> sends changeAOVRequest and prints the resulting active AOV state as JSON.

Expose only render vars that map to real full-resolution image data.

stream

Continuously write frames to stdout for video capture or external tools.

ovusd-shm stream --format raw > frames.bgra
ovusd-shm stream --format png > frames.pngstream

Default to raw output for predictable throughput. Handle SIGINT and SIGTERM by closing the SHM client cleanly.

Common Options

All commands accept:

  • --stream-name <name>: SHM stream name. Default to OV_SHM_STREAM_NAME or the app default such as ovrtx-viewer.

  • --timeout <ms>: operation timeout in milliseconds. Default to 15000.

Timeouts should apply to stream attachment, frame waits, request/response round-trips, and --wait-select unless the command documents a shorter selection-specific timeout.

Building From Source

Build the shared SHM client against the selected SDK headers and libraries, then build the CLI using the generated project's scripts. Share the client implementation with desktop and test clients, while building each native addon for its host runtime ABI. Follow the SDK's loader setup through dependencies and package the required runtime libraries with the app.

The generated local SHM client module should provide:

  • TypeScript types for frames, prim nodes, prim properties, input events, AOV state, and request options

  • a ShmViewerClient class that owns its connection and frame pump

  • PNG encoding without requiring browser APIs

  • optional JPEG encoding through a Node dependency such as sharp

  • native addon loading from installed and source-tree locations

  • a clear error when libovstream.so or libovstream_shm_client.so cannot be loaded

The native addon should:

  • wrap libovstream_shm_client.so

  • build with node-gyp

  • expose blocking frame waits through a safe JS API for CLI use

  • validate frame dimensions, pitch, format, and byte lengths

  • close native handles on process exit or thrown errors

  • avoid exposing raw pointers or file descriptors to JavaScript

Playwright And E2E Testing

Use the headless CLI as the renderer/test backend when Playwright is responsible for UI checks or orchestration:

import { execFile } from 'node:child_process';
import { promisify } from 'node:util';
import { expect, test } from '@playwright/test';

const run = promisify(execFile);
const streamName = process.env.OV_SHM_STREAM_NAME ?? 'ovrtx-viewer';

test('selects a prim through the OVRTX backend', async () => {
  await run('ovusd-shm', ['frame', '--output', 'before.png', '--stream-name', streamName]);
  const { stdout } = await run('ovusd-shm', [
    'click', '0.50', '0.50', '--wait-select', '--stream-name', streamName,
  ]);
  expect(stdout.trim()).toMatch(/^\/World\//);
  await run('ovusd-shm', ['frame', '--output', 'after.png', '--stream-name', streamName]);
});

Testing guidance:

  • start the Python server as a fixture and wait for shmReady before running CLI commands

  • use unique stream names per parallel worker

  • treat the CLI as the source of renderer truth for frame capture, selection, hierarchy, properties, and AOV state

  • keep browser DOM assertions separate from renderer-state assertions

  • preserve captured frames as test artifacts on failure

Example Automation Script

#!/usr/bin/env bash
set -euo pipefail

STREAM_NAME="${OV_SHM_STREAM_NAME:-ovrtx-viewer}"
TARGET="${1:-/World/Cube}"

ovusd-shm info --stream-name "$STREAM_NAME"
ovusd-shm frame --stream-name "$STREAM_NAME" --output before.png

picked="$(ovusd-shm pick --stream-name "$STREAM_NAME" 0.50 0.50)"
echo "picked=${picked}"

ovusd-shm select --stream-name "$STREAM_NAME" "$TARGET"
selected="$(ovusd-shm click --stream-name "$STREAM_NAME" 0.50 0.50 --wait-select)"
test -n "$selected"
test "$selected" != "null"

ovusd-shm props --stream-name "$STREAM_NAME" "$selected" --json > selected-props.json
ovusd-shm frame --stream-name "$STREAM_NAME" --output after.png

This pattern captures a baseline frame, performs a pick/select interaction, verifies that selection state is observable, records selected prim properties, and captures a post-selection frame for visual diffing.

Gotchas

  • LD_LIBRARY_PATH must be set before Node starts. Changing it inside the process is too late for native dynamic loading.

  • --stream-name must exactly match the server --shm-stream-name.

  • Coordinates are 0..1 fractional viewport coordinates, not CSS pixels. The client converts them to render-product pixels after reading a frame.

  • Use the fixed server render resolution for input mapping. Do not use browser DOM size or Electron window size in the CLI.

  • Wait for at least one frame before sending fractional input so width and height are known.

  • Keep JSON control messages small. Do not send frame bytes through JSON.

  • Server callbacks should enqueue work; only the server render owner should call renderer.step(), reset/load scenes, or write live attributes.

  • Raw frame streams need out-of-band width, height, format, and pitch metadata. Capture info alongside raw artifacts.

  • In CI, use unique stream names and clean up stale POSIX SHM segments owned by the current test run only.

Requirements and compatibility

Works with: Claude Code, Claude apps, OpenAI Codex, GitHub Copilot, Cursor, Gemini CLI, Kiro, OpenCode, Goose, Amp, Roo Code, JetBrains Junie, Factory Droid, Devin (ex-Windsurf), Mistral Vibe, OpenHands, Letta

“Works with” means the tool reads the open Agent Skills format; a skill written for one tool may use features others lack.

Needs: Needs Python, Needs Node.js

The author says: Orchestrator skill. Downstream focused references may require NVIDIA GPUs, ovrtx, ovstream, ovui, OpenUSD, Python, Node/React, Tauri, Electron, C++, or cloud GPU deployment access depending on the selected viewer path.

Security

What our scanner noticed· Medium risk

  • Large base64-encoded blobs
  • Instructions that use the network

A static scan is not a guarantee. Skills are third-party content: review a skill before you let an agent use it.

Attribution and licence

Created by
NVIDIA Omniverse
Publisher
NVIDIA (Official)
Source
NVIDIA/skills · skills/omniverse-realtime-viewer · commit fce3cca
Licence
Apache License 2.0 (redistribution allowed)
  • LICENSE (added by AutoGineer from LICENSE-APACHE)
  • ATTRIBUTION.md (written by AutoGineer)
Synced

AutoGineer is not affiliated with or endorsed by NVIDIA Omniverse or NVIDIA.

Download ZIP

The ZIP holds the skill’s files, its licence files and an ATTRIBUTION.md. Download omniverse-realtime-viewer.zip