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://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/cae-cfd-visualization/temporal-playback.md

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

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

Transient / Temporal Playback (time-varying CAE)

Triggers

Use this reference to play back a time-varying (transient) CAE dataset — a timeline with play / scrub, per-frame field interpolation, and a locked-or-auto colormap domain. Trigger terms: transient, temporal, time-varying, timeline, playback, scrub, play/pause, Space to play, per-frame update, field interpolation, "Enable Field Interpolation", keyframe lerp, deforming mesh over time, bumper-beam crash, 51 steps, timeSamples, update_from_usd_time, domain lock across timesteps, "animate the strain field".

Scope: driving a persistent mesh through a series of frames. Reading the native .case/.cgns series into per-step arrays is in cae-data-ingestion.md; the data-plane write mechanics are in driving-cae-viz-via-ovstage.md. This doc is the playback engine that sits between them.

Implement two small application-owned components: a TransientClip containing shared topology plus per-frame arrays, and a PlaybackController that maps play/pause, rate, and scrub input to a frame pair and interpolation factor. Validate with a synthetic two-frame fixture whose midpoint values are known, including both locked and auto colormap domains.

Two engines — pick by how the frames were produced

EngineWhen to usePublication
Per-frame data-plane ARRAY writes (recommended, live)Live server driving frames on demand; stable topology; frames produced at runtimeChanged arrays at one committed ordinal
USD timeSamples (population.update_from_usd_time)All frames baked into a USDA up front; exporting a self-contained clip; not a live serverWhole clip baked into the stage before playback

The examples below use fixed topology and vary points + color per frame. The data-plane path is the live engine; the timeSamples path is for baked-clip export, not the interactive loop.

Per-frame array writes for a fixed-topology clip

For this example, verify that connectivity and node ordering stay fixed across the clip. Author a persistent UsdGeomMesh once, then update points and primvars:displayColor for each sample. Its unchanged faceVertexCounts and faceVertexIndices remain cached.

author persistent UsdGeomMesh ONCE:
    faceVertexCounts / faceVertexIndices   (topology, from frame 0)
    points          (frame 0)              (interpolation implicit)
    primvars:displayColor (frame 0)        (vertex OR uniform — see below)

per timeline tick at fractional t:
    prepare points and colors for the requested sample
    owner writes both arrays at ordinal N and waits for completion
    advance_write_floor(N).wait()  ->  renderer.step(ordinal=N)

Use the typed-write and tensor-lifetime mechanics in references/ovstage-data-plane/README.md. The publication choice for meshes, curves and changing topology is defined in driving-cae-viz-via-ovstage.md.

Stable topology and changing geometry

Cache face counts/indices when connectivity and node ordering are unchanged. Equal point counts alone do not establish that condition. For changing mesh topology, update face counts/indices with points and every affected primvar on the existing Mesh, in one committed publication. Replace structure when prim type or composition changes; follow control-to-publication routing.

Interpolate samples only when topology, ordering and field association match. Otherwise select a complete sample instead of blending unrelated vertices.

Reusing an OVStage display prim does not establish that scientific input caches remain valid across samples. Before rerunning operators on changed solver data, follow the selected release's ownership/cache contract, reached through references/dependencies/README.md → Warp SimData → datasets and fields.

Field interpolation = numpy lerp between keyframes

Field interpolation is a plain linear interpolation of both points and field between the two bracketing keyframes at a fractional t, then recolor. No GPU, no operator recompute:

i0 = int(np.floor(t)); i1 = min(i0 + 1, n_frames - 1); a = t - i0
pts = (1 - a) * load_points(i0) + a * load_points(i1)   # deformed geometry
fld = (1 - a) * load_field(i0)  + a * load_field(i1)    # interpolated scalar
colors = apply_lut(fld, vmin, vmax, lut)                # recolor

With interpolation off, snap to round(t) and use that frame's arrays verbatim. This is a pure CPU coloring/geometry choice made just before the data-plane write.

Domain LOCK is a pure CPU coloring choice (not a stage-wide setting)

Domain locking freezes the colormap range so it does not auto-rescale per frame. In this engine it is simply which (vmin, vmax) you pass to the LUT before the write:

  • Locked: use the clip's global field range (min/max across all frames). The color of a given strain value is stable through the whole animation — the correct default for reading how a field grows.

  • Auto: use the current frame's min/max. Each frame fills the colormap but values are not comparable across time.

if lock_domain:
    vmin, vmax = clip.field_global_range()     # computed once, cached
else:
    vmin, vmax = float(np.nanmin(fld)), float(np.nanmax(fld))

Compute the global range once and cache it (a full sweep of the clip). It is not a per-attribute USD metadata flag; in a bake-per-vertex-color viewer it is just the range argument.

Architecture: TransientClip + PlaybackController

  • TransientClip — random access over a stable topology: n_frames, node_count, the authored-once face_counts / face_indices, lazy load_points(i) / load_field(i) (so a long clip is never fully resident), the field_kind ("node" → per-vertex color, "element" → per-face color), and field_global_range() for the lock. Build one from the native readers in cae-data-ingestion.md, loading geometry and fields at the same source time code and verifying topology before reuse.

  • PlaybackController — turns a fractional t into (points, displayColor, (vmin,vmax)) ready for the write, applying interpolation and the locked-or-auto domain. Config is cheap: set_colormap / set_lock / set_interpolate.

Playback state and the render loop:

  • New command kinds on the serialized owner: timeline{t} (scrub), play{playing,fps,speed}, domain_lock{locked}, interpolate{enabled} — all enqueued like any other control (see driving-cae-viz-via-ovstage.md).

  • While playing, advance t += dt * fps * speed each tick, looped; scrub sets t directly.

  • Everything runs under one ordinal / one advance_write_floor on the render thread, so transient playback composes with an emitter or camera move in the same tick.

The USD timeSamples alternative (baked clips only)

For a self-contained exported clip, author every frame as USD timeSamples and drive it with population.update_from_usd_time(stage, ordinal=N, time=seconds) — time is in seconds (via the stage timeCodesPerSecond), then advance_write_floor(N) + step. Use this only when all frames are baked up front (clip export), not for the live server — baking a long transient into a USDA is wasteful compared to streaming two arrays per tick.

ovstage inline-USDA parser is stricter than pxr: no semicolon one-line blocks — one attribute per line. A timeSamples block that pxr accepts on one line will fail to parse through open_usd_from_string. Author each sample on its own line.

Gotchas

  • Cache unchanged topology. Publish changed connectivity and all dependent arrays together when a sample changes mesh topology; count changes alone do not require a new prim.

  • field_kind picks the primvar interpolation — element fields color per face (uniform, length F), node fields per vertex (vertex, length N). Author the matching interpolation once; the per-frame color array length must match it.

  • Lock = global range; auto = per-frame range. Default to lock so a value's color is comparable across time; compute the global range once and cache it.

  • Don't bake a live transient into timeSamples. Data-plane array writes are the live engine; timeSamples is for baked-clip export, and its inline USDA must be one attribute per line.

See also

  • cae-data-ingestion.md — reading the .case/.cgns series into the per-step arrays a TransientClip loads.

  • driving-cae-viz-via-ovstage.md — the data-plane array-publication flow, dependent primvar lengths, and the serialized command loop.

  • usd-authoring-and-materials.md — primvar interpolation (vertex vs uniform) and colormap authoring.

  • streaming-cae-viewer.md — the timeline UI + sceneState timeline contract that surfaces this to the browser.

  • references/ovstage-population/README.md — update_from_usd_time for the baked-clip path.

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