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/object-selection/README.md

Object Selection

Triggers

Use this skill for pick objects, click to select, object selection, native picking, pick queries, marquee selection, pickable prims, select prim, or wrong prim selected requests.

Use this for viewport picking and selected-prim state. Selection visuals live in selection-feedback; selected-prim properties live in prim-info-display.

For ovrtx, the recommended path is the native picking API documented by native-picking-selection.

For ovrtx selection or picking behavior not covered here, read references/dependencies for acquisition guidance and supplemental dependency documentation.

Read viewer-input-routing first when selection is driven by viewport clicks, marquee gestures, WebRTC input, or click-vs-drag discrimination.

Do not create new GpuPicker, cpu_picking.py, seg_outline.py, Warp outline, learn_mapping, isolation discovery, ID drift repair, or CPU ray/AABB fallback systems for generated apps.

Native Picking Facts

  • Queue pick work with Renderer.enqueue_pick_query() in Python or ovrtx_enqueue_pick_query() in C/C++.

  • Queue the pick query before the renderer.step() that should produce the result.

  • Convert RenderProduct pixels to top-left normalized coordinates before enqueueing. left and top are inclusive; use one pixel past the selected bounds for right and bottom.

  • The consumed step exposes the synthetic render var ovrtx.OVRTX_RENDER_VAR_PICK_HIT / OVRTX_RENDER_VAR_PICK_HIT.

  • Pick hits store prim path IDs. Resolve them with renderer.resolve_prim_path_id() in Python or the C path dictionary utilities before printing names or updating selection.

  • Multiple pick queries for the same RenderProduct before a single step are not queued independently; treat the last query as authoritative.

  • The synthetic pick-hit output is automatic for the consuming step; do not add ovrtx_pick_hit to USD orderedVars.

Pickability

Mark selectable prims with the native pickable helper. Pickability is renderer interaction state, so the native OVRTX helper remains the right surface here. Pass renderer path IDs, such as decoded pick-hit IDs; do not write a legacy pickable attribute through renderer.write_attribute(). New apps should not build or maintain a separate segmentation ID mapping just to decide what can be selected.

Python:

def set_pickable(renderer, prim_path_ids: list[int], enabled: bool) -> None:
    if prim_path_ids:
        renderer.set_pickable(prim_path_ids, enabled)

C/C++:

Pass one boolean per prim path. This two-prim example uses path IDs already resolved by the same renderer:

enum { prim_path_count = 2 };
const ovx_primpath_t prim_path_ids[prim_path_count] = {
    first_prim_path_id, second_prim_path_id
};
const bool pickable[prim_path_count] = {true, true};
ovrtx_set_pickable(renderer, prim_path_ids, prim_path_count, pickable);

Keep both arrays alive until the queued operation completes.

When a frontend sends makePrimsSelectable or makePrimsPickable, update the server's canonical pickable set and write pickability for the changed prims. Do not recompute CPU bbox maps or segmentation ID maps as part of the normal ovrtx selection path.

Click Pick Flow

Only run selection on a completed click gesture. Camera orbit/pan/zoom drags must not also select.

def pick_at_render_pixel(renderer, render_product_path: str, x: int, y: int, width: int, height: int, ordinal: int) -> list[str]:
    renderer.enqueue_pick_query(
        render_product_path=render_product_path,
        left_ndc=x / width,
        top_ndc=y / height,
        right_ndc=(x + 1) / width,
        bottom_ndc=(y + 1) / height,
    )

    products = renderer.step(
        render_products={render_product_path},
        delta_time=1.0 / 60.0,
        ordinal=ordinal,
    )

    frame = products[render_product_path].frames[0]
    return resolve_pick_hit_paths(renderer, frame)

Keep UI coordinate conversion outside the picker:

  1. Measure the displayed video/image rectangle.

  2. Reject clicks in letterboxed areas.

  3. Convert to RenderProduct pixel coordinates.

  4. Clamp to [0, width - 1] and [0, height - 1].

  5. Call the native pick query.

Marquee Selection

Drag selection uses the same native API with a larger rectangle:

def marquee_pick(renderer, render_product_path: str, x0: int, y0: int, x1: int, y1: int, width: int, height: int, ordinal: int) -> list[str]:
    left = min(x0, x1)
    top = min(y0, y1)
    right = max(x0, x1) + 1
    bottom = max(y0, y1) + 1

    renderer.enqueue_pick_query(
        render_product_path=render_product_path,
        left_ndc=left / width,
        top_ndc=top / height,
        right_ndc=right / width,
        bottom_ndc=bottom / height,
    )

    products = renderer.step(
        render_products={render_product_path},
        delta_time=1.0 / 60.0,
        ordinal=ordinal,
    )

    frame = products[render_product_path].frames[0]
    return resolve_pick_hit_paths(renderer, frame)

For additive or subtractive marquee modes, apply modifier keys after resolving the hit paths and before broadcasting the canonical selection state.

In attached-stage viewers, these helpers run inside the sole render owner. Pass the latest committed ordinal already chosen for that frame; a picker must not call an un-ordinaled or competing renderer.step().

Decode Pick Hits

import numpy as np
import ovrtx


def resolve_pick_hit_paths(renderer: ovrtx.Renderer, frame) -> list[str]:
    if ovrtx.OVRTX_RENDER_VAR_PICK_HIT not in frame.render_vars:
        return []
    pick_var = frame.render_vars[ovrtx.OVRTX_RENDER_VAR_PICK_HIT]

    mapping = pick_var.map(device=ovrtx.Device.CPU)
    try:
        magic = int(np.from_dlpack(mapping.params["magic"]).reshape(-1)[0])
        version = int(np.from_dlpack(mapping.params["version"]).reshape(-1)[0])
        hit_count = int(np.from_dlpack(mapping.params["hitCount"]).reshape(-1)[0])
        prim_path_ids = np.from_dlpack(mapping["primPath"]).copy().reshape(-1)
    finally:
        mapping.unmap()

    if magic != ovrtx.OVRTX_PICK_HIT_MAGIC or version != ovrtx.OVRTX_PICK_HIT_VERSION:
        raise RuntimeError("Unexpected ovrtx pick-hit schema")

    paths: list[str] = []
    seen: set[str] = set()
    for prim_path_id in prim_path_ids[:hit_count]:
        path = renderer.resolve_prim_path_id(int(prim_path_id))
        if path and path not in seen:
            paths.append(path)
            seen.add(path)
    return paths

If the app needs world-space hit data for gizmos or labels, follow native pick decoding for the required query flag and validated result access.

Selection State

Keep selection state centralized on the renderer/server side:

def select_paths(paths: list[str], mode: str = "replace") -> None:
    previous = set(selection.paths)

    if mode == "add":
        current = previous | set(paths)
    elif mode == "subtract":
        current = previous - set(paths)
    else:
        current = set(paths)

    selection.paths = sorted(current)
    selection.mesh_paths = expand_to_descendant_meshes(selection.paths)
    selection_feedback.update(selection.mesh_paths)
    message_handler.send_message("stageSelectionChanged", {"prims": selection.paths})

Tree/sidebar selection and viewport selection must both call the same state transition. The frontend mirrors stageSelectionChanged; it should not maintain an independent authoritative selection.

For Xform or Scope selection, use stage-hierarchy to expand to descendant mesh paths only for visual feedback. Preserve the user's selected tree path for the stage tree and info panel.

Runtime Feature Handoff

Selection is not a blanket OVStage write. Keep canonical selected paths in the app/runtime controller, use native OVRTX selection groups for outlines, and only handoff to OVStage when selection drives runtime data such as transforms, material/effect attributes, visibility, physics poses, or reversible viewer-owned state.

native pick result
  -> resolve prim path IDs for the active stage generation
  -> update canonical selected paths
  -> update native OVRTX selection groups/outlines
  -> dispatch feature commands to the OVStage runtime owner
  -> feature writes publish at ordinal N
  -> attached renderer consumes N on its normal render step

Do not let the picker callback perform direct OVRTX material, visibility, or omni:xform writes. The picker should enqueue intent, and the single runtime owner should serialize OVStage writes with the same ordinal publication protocol used for camera, animation, transform tools, and physics handoff.

Scene Lifecycle

On scene switch or renderer reset:

  • Do not let pick queries run while the renderer is loading or resetting.

  • Clear selected paths, hover state, info panels, and pending pick requests.

  • Reapply pickability after the new stage has loaded.

  • Clear previous native selection outline groups through selection-feedback.

  • Do not preserve pick-hit records across scenes.

Deprecated Segmentation Fallback

Segmentation-buffer picking from InstanceSegmentationSD is a deprecated compatibility path for older OVRTX builds or custom post-process tools. It is not the recommended ovrtx path.

Only use the deprecated path when native pick queries are unavailable and the user explicitly accepts the limitations: per-frame ID buffers, scene-local IDs, ID-to-path mapping, ID drift after reloads, and possible mismatch with UI selection. Do not scaffold it in new ovrtx apps.

Gotchas

  • Selection bugs are usually coordinate bugs. Convert through the visible image/video rectangle before calling the native pick query.

  • Pick queries are consumed by a renderer step; enqueue before stepping.

  • Native picking returns path IDs, not path strings.

  • Write pickability before expecting a prim to appear in pick results.

  • Do not call renderer.step() concurrently with scene reset/load or from input callbacks. Enqueue selection work for the render loop.

  • A selection outline requires selection-feedback; picking only decides what is selected.

See also: viewer-input-routing, native-picking-selection, selection-feedback, prim-info-display, camera-controls, local-viewer, streaming-server, streaming-messages, stage-hierarchy, stage-management.

Generated Module Checklist - selection_controller.py

  • [ ] Converts UI coordinates to RenderProduct pixels before picking.

  • [ ] Queues click picks with Renderer.enqueue_pick_query() using normalized RenderProduct coordinates.

  • [ ] Queues marquee picks with a native pick rectangle.

  • [ ] Decodes OVRTX_RENDER_VAR_PICK_HIT.

  • [ ] Resolves prim path IDs before broadcasting selection.

  • [ ] Writes pickability with Renderer.set_pickable() or ovrtx_set_pickable() using renderer path IDs.

  • [ ] Keeps selected paths server/runtime authoritative.

  • [ ] Expands selected Xforms/Scopes to descendant mesh paths only for feedback.

  • [ ] Dispatches transform, effect, visibility, and physics side effects through the OVStage runtime owner instead of direct picker writes.

  • [ ] Clears selection and pending pick state on scene switch.

  • [ ] Does not create GpuPicker, cpu_picking.py, seg_outline.py, or Warp outline systems.

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