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

Omniverse Realtime Viewer

OpenAIOfficialCreated by NVIDIA OmniverseDistributed by OpenAI

  • Skill
  • Meta (authoring, orchestration, planning, memory)
  • Software development
  • Apache-2.0
  • Medium risk

From the author

Use as the top-level router for Omniverse Realtime Viewer USD app requests 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 orchestrator for developing USD-based viewer applications. It provides a structured approach to selecting implementation recipes, enforcing architectural conventions, and managing dependencies. By utilizing a set of focused reference documents, it guides developers through the lifecycle of building, configuring, and validating viewer applications, ensuring consistent behavior across different delivery paths such as local workstations or streaming environments.

What it does

  • Routes development requests to specific implementation recipes based on the target delivery path.
  • Enforces architectural rules for USD rendering, stage management, and UI integration.
  • Provides guidance on camera selection, input handling, and viewport interaction.
  • Standardizes the integration of AOV switching and render product configuration.
  • Facilitates the capture of validation evidence for viewer readiness.
  • Maintains separation between user-provided USD files and application-specific session state.

When to use it

  • When initiating the development of a new USD viewer application within the Omniverse ecosystem.
  • When determining the appropriate delivery path for a viewer, such as local native or streaming.
  • When needing to implement consistent camera, selection, or UI behavior across viewer projects.
  • When integrating advanced rendering features like AOV switching into an existing viewer.

Advantages

  • Ensures architectural consistency across diverse viewer implementations.
  • Provides a clear, modular path for adding specific capabilities like input routing or data display.
  • Promotes clean separation of concerns between stage data and viewer-owned state.
  • Standardizes the validation process for consistent delivery.

Usage examples

“I need to build a browser-based viewer for USD files.”
The agent will route you to the streaming viewer recipe, providing guidance on WebRTC stream setup, UI integration, and viewport interaction.
“How do I implement AOV switching in my local USD viewer?”
The agent will provide the AOV switching reference, detailing how to map render vars and update the display state on the server.

Install

  1. Download the skill as a ZIP:

    https://autogineer.com/agent-skills/openai/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

This is the top-level entry point for the Omniverse Realtime Viewer skill package. It is self-contained: all required routing, conventions, and validation guidance live in the selected references.

Use the focused reference documents as implementation recipes. This file chooses the right recipes and preserves the architectural rules that must hold across all generated viewer apps.

Instructions

Start by classifying the requested viewer, then read only the references needed for that delivery path and feature set. Implement the render path first, layer interaction and UI behavior on top of it, and finish by capturing validation evidence from references/validation.md.

Read Order

  1. Read references/routing.md to choose the delivery path and focused references.

  2. Read references/conventions.md before implementing camera, input, selection, viewport, streaming protocol, scene loading, or environment behavior.

  3. For broad viewer requests, read references/usd-viewer-app/README.md.

  4. If the delivery path is unclear, read references/streaming-vs-local/README.md.

  5. If the prompt includes layout, panels, controls, inspectors, status, or UX, read references/viewer-ux-workflow/README.md and then the focused viewer UI references. This applies to React/WebRTC, Tauri, Electron, ovui, ovwidgets, and Dear ImGui apps; "frontend" means user-facing UI, not only browser UI.

  6. For viewport interaction, read references/viewer-input-routing/README.md before references/camera-controls/README.md, references/native-picking-selection/README.md, or references/object-selection/README.md.

  7. Read only the focused capability references needed for the requested app.

  8. Use references/validation.md to capture review evidence before handoff.

Non-Negotiables

  • Use ovrtx for all USD and 3D rendering.

  • Browser apps display an ovstream WebRTC video stream plus UI. The browser does not render USD geometry.

  • Do not substitute WebGL, Three.js, Babylon.js, PlayCanvas, A-Frame, model-viewer, react-three-fiber, glTF browser viewers, or other client-side 3D renderers.

  • If local validation cannot run because the GPU/runtime environment is absent, scaffold the ovrtx path and document the runtime requirement. Do not add a browser-renderer fallback.

  • Keep user USD files unmodified. Viewer cameras, render products, render vars, settings, selection metadata, and runtime state belong in session/composite layers or app state.

  • Keep one owner for renderer.step(), stage mutation, native picking, selection writes, and live attribute writes.

  • Keep dependency acquisition in references/dependencies/README.md and deployment choices in references/cloud-deployment/README.md; do not duplicate package locations or deployment setup.

Focused Reference Families

  • Entry points and recipes: references/usd-viewer-app/README.md, references/streaming-viewer-recipe/README.md, references/ovui-local-viewer-recipe/README.md, references/streaming-vs-local/README.md, references/electron-shm-viewer/README.md, references/ovwidgets-editor-shell/README.md.

  • Rendering and stage: references/ovrtx-rendering/README.md, references/stage-loading/README.md, references/stage-management/README.md, references/render-settings/README.md, references/aov-switching/README.md, references/stage-hierarchy/README.md, references/stage-queries/README.md, references/stage-attribute-reads/README.md, references/prim-transform-safety/README.md, references/usd-sample-data/README.md.

  • Delivery and runtime: references/streaming-server/README.md, references/streaming-client/README.md, references/streaming-messages/README.md, references/streaming-lifecycle/README.md, references/local-viewer/README.md, references/tauri-local-viewer/README.md, references/cpp-native-viewer/README.md, references/headless-shm-cli/README.md, references/viewer-backend-interface/README.md, references/webgl-shm-transport/README.md.

  • Viewer UI/UX: references/viewer-ux-workflow/README.md, references/viewer-layout-patterns/README.md, references/viewer-control-patterns/README.md, references/viewer-data-view-patterns/README.md, references/viewer-feedback-status/README.md.

  • Interaction: references/viewer-input-routing/README.md, references/camera-controls/README.md, references/object-selection/README.md, references/native-picking-selection/README.md, references/selection-feedback/README.md, references/selection-animation/README.md, references/transform-manipulator/README.md, references/gl-viewport-overlay/README.md, references/ovui-library/README.md, references/prim-pick-effects/README.md, references/prim-info-display/README.md, references/viewport-overlays/README.md.

  • Infrastructure: references/dependencies/README.md, references/windows-native-setup/README.md, references/cloud-assets/README.md, references/cloud-deployment/README.md, references/troubleshooting/README.md.

Build Workflow

  1. Classify the prompt by delivery path, target user, required capabilities, runtime environment, validation needs, and explicit constraints.

  2. Select a small reference set. Start with the recipe or routing reference, then add focused capabilities such as camera, picking, hierarchy, properties, render settings, transform tools, cloud assets, or deployment.

  3. Read selected references before writing app code. Follow their build order, import order, data-channel contracts, and renderer ownership rules.

  4. Implement the core render path first, then input routing and camera, then selection and data panels, then scene/settings features, then packaging or deployment.

  5. Treat the selected references as the behavior contract for API shape, compatibility, and generated project structure.

  6. Capture validation evidence before calling the viewer ready.

Examples

  • For a browser viewer request, use the streaming recipe references plus camera, picking, hierarchy, properties, render settings, and stream-status references.

  • For a local workstation viewer request, use the local or native delivery references plus renderer setup, stage loading, viewport input, and validation.

Completion Checklist

  • Selected references match the user's intent and delivery path.

  • No code path uses a browser-side 3D renderer for USD.

  • The generated app has one clear owner for render stepping and stage mutation.

  • User USD files remain untouched by viewer-owned session data.

  • Camera, input, selection, scene loading, and stream behavior follow references/conventions.md.

  • Setup/build/run results and visual interaction evidence are captured with references/validation.md.

Files

references/stage-loading/README.md

Stage Loading

Triggers

Use this skill for load USD, RenderProduct, RenderVar, black frame, composite stage, session layer, camera aspect, open_usd, add_usd_reference, or Unable to find RenderProduct prim.

ovrtx needs a complete render pipeline in the stage: Camera -> RenderProduct -> RenderVar -> RenderSettings. Most user USD files do not include this, so viewers load a generated root/composite stage that sublayers the user scene and authors viewer-owned render prims.

For ovrtx stage composition, render pipeline, or release-specific loading behavior not covered here, read references/dependencies for acquisition guidance and supplemental dependency documentation.

ovrtx 0.3 Stage Composition APIs

Use the explicit 0.3 composition APIs:

  • renderer.open_usd(path) opens a file/URL as the active root layer. Calling it again replaces the previous root layer.

  • renderer.open_usd_from_string(usda) opens generated inline USDA as the active root layer. Use this for viewer/session USD that sublayers a user scene and adds cameras, RenderProducts, RenderVars, and RenderSettings.

  • renderer.add_usd_reference(path, prefix_path="/SomePrim") adds referenced content under a prim path after a root stage is already open.

  • renderer.add_usd_reference_from_string(usda, prefix_path="/SomePrim") is the inline-string equivalent for additive referenced content. Inline referenced layers need a defaultPrim.

  • renderer.remove_usd(handle) removes content added by add_usd_reference*.

  • renderer.reset_stage() clears the runtime stage to empty. It is not required before normal scene replacement because open_usd* already replaces the active root layer.

Do not use older implicit stage-addition or anonymous-layer staging patterns for 0.3 stage loading.

Generated Root Stage Pattern

For local and streamed Omniverse Realtime Viewers, generate one root USDA layer that sublayers the user scene and contains only viewer camera, render product, render vars, and render settings. Do not inject lights here unless the user asked for viewer-controlled lighting and the app exposes a verified lighting capability or explicit reload/profile workflow.

CAMERA_PATH = "/Session/Cameras/Main"
RENDER_PRODUCT_PATH = "/Session/Render/Viewport"

def viewer_root_usda(scene_path: str, width: int, height: int) -> str:
    h_aperture = 20.955
    v_aperture = h_aperture * float(height) / float(width)
    scene_ref = scene_path.replace("\\", "/")
    return f"""#usda 1.0
(
    subLayers = [
        @{scene_ref}@
    ]
    defaultPrim = "Session"
)
def Scope "Session" {{
    def Scope "Cameras" {{ def Camera "Main" {{
        float focalLength = 18.15
        float horizontalAperture = {h_aperture}
        float verticalAperture = {v_aperture}
        float2 clippingRange = (0.01, 10000000)
        token projection = "perspective"
        matrix4d xformOp:transform = ((1,0,0,0),(0,1,0,0),(0,0,1,0),(0,0,0,1))
        uniform token[] xformOpOrder = ["xformOp:transform"]
    }} }}
    def Scope "Render" {{ def RenderProduct "Viewport" {{
        rel camera = </Session/Cameras/Main>
        rel orderedVars = [</Session/Render/Vars/LdrColor>, </Session/Render/Vars/InstanceSeg>]
        uniform int2 resolution = ({int(width)}, {int(height)})
    }}
    def Scope "Vars" {{
        def RenderVar "LdrColor"
        {{
            uniform string sourceName = "LdrColor"
        }}
        def RenderVar "InstanceSeg"
        {{
            uniform string sourceName = "InstanceSegmentationSD"
        }}
    }} }}
}}
"""
renderer.open_usd_from_string(viewer_root_usda(str(stage_path), width, height))
products = renderer.step(render_products={RENDER_PRODUCT_PATH}, delta_time=1 / 60)
with products as ctx:
    product = ctx[RENDER_PRODUCT_PATH]

Direct Frame Validation

For local desktop viewers, always separate renderer validation from native UI presentation validation. After opening the generated root stage, step the same RenderProduct path the viewport will use and save a direct LdrColor artifact before debugging the window.

A nonblank direct LdrColor frame proves that the generated Camera -> RenderProduct -> RenderVar wiring is basically working. If the native window is still black or blank after that, the next suspect is the ovui presentation path, not the render product path, camera relation, or USD composition.

If the direct frame is blank, continue debugging this skill's concerns: user sublayer path, camera path, render product path, render var source name, resolution, camera transform, stage lighting, material/plugin resolution, and load-operation errors.

Composite File Pattern

Streaming servers should prefer a wrapper .usda written beside the user stage. The wrapper sublayers the user scene, injects the server camera/render product/render vars, and is passed to renderer.open_usd(composite_path). During scene switches, the next open_usd() call replaces the previous root stage; do not reset first unless the user explicitly requested an empty stage.

The reference streaming server uses camera path /OVCamera and render product path /Render/OVServer/ViewportTexture0.

OV_CAMERA_PRIM = "/OVCamera"
OV_RENDER_PRODUCT = "/Render/OVServer/ViewportTexture0"
CAMERA_HORIZONTAL_APERTURE = 20.955

def make_composite_stage(scene_url: str, width=1920, height=1080) -> str:
    scene_ref = scene_url.replace("\\", "/")
    safe_width = max(1, int(width))
    safe_height = max(1, int(height))
    vertical_aperture = CAMERA_HORIZONTAL_APERTURE * float(safe_height) / float(safe_width)
    return f'''#usda 1.0
(
    subLayers = [
        @{scene_ref}@
    ]
)

def Camera "OVCamera"
{{
    float2 clippingRange = (1, 10000000)
    float focalLength = 18.15
    float horizontalAperture = {CAMERA_HORIZONTAL_APERTURE:.3f}
    float verticalAperture = {vertical_aperture:.4f}
    token projection = "perspective"
    double3 xformOp:translate = (-553.5, 246.6, -22.5)
    uniform token[] xformOpOrder = ["xformOp:translate"]
}}

def "Render"
{{
    def "OVServer"
    {{
        def RenderProduct "ViewportTexture0" (
            prepend apiSchemas = ["OmniRtxSettingsCommonAdvancedAPI_1", "OmniRtxSettingsPtAdvancedAPI_1", "OmniRtxSettingsRtAdvancedAPI_1"]
        )
        {{
            token omni:rtx:rendermode = "RealTimePathTracing"
            bool omni:rtx:pt:diAOV = 1
            bool omni:rtx:pt:giAOV = 1
            bool omni:rtx:pt:diffuseFilterAOV = 1
            bool omni:rtx:pt:reflectionsAOV = 1
            bool omni:rtx:pt:refractionFilterAOV = 1
            bool omni:rtx:pt:refractionsAOV = 1
            bool omni:rtx:pt:selfIllumAOV = 1
            bool omni:rtx:pt:volumesAOV = 1
            bool omni:rtx:pt:worldNormalsAOV = 1
            bool omni:rtx:pt:worldPosAOV = 1
            bool omni:rtx:pt:zDepthAOV = 1
            bool omni:rtx:pt:denoising:optix:denoiseAOVs = 1
            float omni:rtx:pt:zDepthMin = 0.1
            float omni:rtx:pt:zDepthMax = 10000
            int omni:rtx:pt:maxSamplesPerLaunch = 2073600
            float omni:rtx:rtpt:modulatingRoughnessThreshold = 0.08
            rel camera = <{OV_CAMERA_PRIM}>
            rel orderedVars = [
                </Render/Vars/LdrColor>,
                </Render/Vars/HdrColor>,
                </Render/Vars/Depth>,
                </Render/Vars/Normal>,
                </Render/Vars/InstanceSeg>,
                </Render/Vars/SemanticSeg>,
                </Render/Vars/Metallic>,
                </Render/Vars/Roughness>,
                </Render/Vars/Emissive>,
                </Render/Vars/Diffuse>,
                </Render/Vars/Specular>,
                </Render/Vars/AO>,
                </Render/Vars/DirectDiffuse>,
                </Render/Vars/DirectSpecular>,
                </Render/Vars/IndirectDiffuse>,
                </Render/Vars/IndirectSpecular>,
                </Render/Vars/MotionVectors>,
            ]
            uniform int2 resolution = ({safe_width}, {safe_height})
        }}
    }}

    def "Vars"
    {{
        def RenderVar "LdrColor"
        {{
            uniform string sourceName = "LdrColor"
        }}
        def RenderVar "HdrColor"
        {{
            uniform string sourceName = "HdrColor"
        }}
        def RenderVar "Depth"
        {{
            uniform string sourceName = "DepthSD"
        }}
        def RenderVar "Normal"
        {{
            uniform string sourceName = "NormalSD"
        }}
        def RenderVar "InstanceSeg"
        {{
            uniform string sourceName = "InstanceSegmentationSD"
        }}
        def RenderVar "SemanticSeg"
        {{
            uniform string sourceName = "SemanticSegmentationSD"
        }}
        def RenderVar "Metallic"
        {{
            uniform string sourceName = "Metallic"
        }}
        def RenderVar "Roughness"
        {{
            uniform string sourceName = "Roughness"
        }}
        def RenderVar "Emissive"
        {{
            uniform string sourceName = "Emissive"
        }}
        def RenderVar "Diffuse"
        {{
            uniform string sourceName = "DiffuseAlbedoSD"
        }}
        def RenderVar "Specular"
        {{
            uniform string sourceName = "Specular"
        }}
        def RenderVar "AO"
        {{
            uniform string sourceName = "AmbientOcclusion"
        }}
        def RenderVar "DirectDiffuse"
        {{
            uniform string sourceName = "DirectDiffuse"
        }}
        def RenderVar "DirectSpecular"
        {{
            uniform string sourceName = "DirectSpecular"
        }}
        def RenderVar "IndirectDiffuse"
        {{
            uniform string sourceName = "IndirectDiffuse"
        }}
        def RenderVar "IndirectSpecular"
        {{
            uniform string sourceName = "IndirectSpecular"
        }}
        def RenderVar "MotionVectors"
        {{
            uniform string sourceName = "MotionVectors"
        }}
    }}

    def RenderSettings "OVRenderSettings"
    {{
        rel products = [<{OV_RENDER_PRODUCT}>]
    }}
}}

# Override EffectLayer shaders to disable selection glow.
# In ovrtx, no OmniGraph runtime drives EffectLayerMT.mdl's animation input.
# Setting Fader=0 forces a clean load-time non-highlighted state.
over "World"
{{
    over "Misc"
    {{
        over "Looks"
        {{
            over "Concrete_Rough"
            {{
                over "EffectLayer"
                {{
                    float inputs:Fader = 0
                }}
            }}
            over "Steel_Stainless"
            {{
                over "EffectLayer"
                {{
                    float inputs:Fader = 0
                }}
            }}
            over "MetallicGreen_OmniPbr"
            {{
                over "EffectLayer"
                {{
                    float inputs:Fader = 0
                }}
            }}
        }}
    }}
}}
'''

Pass the injected RenderProduct path to renderer.step().

The EffectLayer override block above is a material-effect example, not a baseline stage-loading requirement. For a general viewer, only generate equivalent over blocks when the active stage actually contains compatible EffectLayer shader paths and the app intends to use material-driven pick effects. Keep those overrides in the composite/session layer before runtime effect writes use PrimMode.EXISTING_ONLY.

The OmniRtx API schemas and path-tracing AOV flags are required viewer-owned render pipeline metadata. Recommended viewer implementations author the schemas on the RenderProduct; if a target ovrtx build expects them on RenderSettings, keep the same schema list and flag values on the render settings prim instead of dropping them.

Do not use inline one-line prim bodies such as def RenderVar "LdrColor" { uniform string sourceName = "LdrColor" } or nested one-line override bodies such as over "EffectLayer" { float inputs:Fader = 0 }. Some ovrtx-bundled USD parser builds reject or misdiagnose these compact forms, especially when generated through Python strings with escaped braces. Use the multi-line brace form shown above for every generated def, over, and nested override block.

Generated USDA Self-Check

Before calling renderer.open_usd() or renderer.open_usd_from_string() with a generated wrapper, validate the exact generated text with OpenUSD in the selected pxr subprocess. This catches malformed braces, bad asset references, and wrong value syntax before the ovrtx process enters its render/load path.

Use this validation for generated app scaffolds and tests. Keep it out of the ovrtx render process when the app otherwise follows the pxr-subprocess isolation contract.

Initial Resolution And Aspect

The RenderProduct resolution and camera aperture must agree. Derive verticalAperture from horizontalAperture * height / width when creating session/composite camera data.

Browser-streamed Omniverse Realtime Viewer apps should use a fixed server render resolution, typically 1920x1080, and let the frontend display the video with object-fit: contain. CSS layout changes should not rebuild session/composite camera data.

Write the composite into the same directory as the user stage and reference the user stage by basename so relative textures, MDL files, and sublayers resolve:

stage_dir = os.path.dirname(os.path.abspath(url))
stage_basename = os.path.basename(url)
stage_stem = os.path.splitext(stage_basename)[0]
composite_path = os.path.join(stage_dir, f"_ovrtx_composite_{stage_stem}.usda")
with open(composite_path, "w", encoding="utf-8") as f:
    f.write(make_composite_stage(stage_basename, width, height))

renderer.open_usd(composite_path)
products = renderer.step(render_products={OV_RENDER_PRODUCT}, delta_time=1 / 60)

Dynamic Scene Root

Do not assume the loaded scene root is /World. Some assets use roots such as /stage, and hardcoded /World paths break hierarchy, selection, and pickable-prim setup.

When opening the USD for metadata, detect and store the root prim path:

  1. Prefer /World if it exists.

  2. Otherwise use stage.GetDefaultPrim() if valid.

  3. Otherwise use the first pseudo-root child.

Pass this root_prim_path through the load result so frontend hierarchy and selection code can query the correct root.

Implementation: pxr_worker subprocess

Do not import pxr (OpenUSD Python) in the main ovrtx process — it conflicts with ovrtx's bundled USD. Run all pxr queries in a separate subprocess:

# pxr_worker.py — runs in subprocess, communicates via JSON over stdin/stdout
def cmd_get_root_prim_path():
    """Detect the best root prim path for the stage."""
    if not _stage:
        return {"ok": False, "error": "no stage loaded"}

    # 1. Prefer /World if it exists
    world = _stage.GetPrimAtPath("/World")
    if world.IsValid():
        return {"ok": True, "path": "/World"}

    # 2. Try DefaultPrim
    default_prim = _stage.GetDefaultPrim()
    if default_prim and default_prim.IsValid():
        return {"ok": True, "path": str(default_prim.GetPath())}

    # 3. First pseudo-root child
    for child in _stage.GetPseudoRoot().GetChildren():
        return {"ok": True, "path": str(child.GetPath())}

    return {"ok": True, "path": "/"}

The main server proxies this via PxrWorkerClient.get_root_prim_path() and caches the result in self.current_stage_root_path after each successful load.

Protocol: root_prim_path in openStageResult

The server includes root_prim_path in every openStageResult and push_initial_state message:

self.send_message("openStageResult", {
    "url": active_url,
    "result": "success",
    "root_prim_path": server.current_stage_root_path,
})
# Send children from detected root, not hardcoded /World
children = server._pxr.get_children(root_path)
self.send_message("getChildrenResult", {
    "prim_path": root_path,
    "children": children,
})

Skip-Reload Optimization

When the frontend sends openStageRequest for a stage that is already loaded (e.g., on reconnect or duplicate requests), skip the expensive renderer reload:

def _stage_path_key(self, url: str) -> str:
    """Normalize path for comparison."""
    return os.path.normcase(os.path.abspath(url))

def _load_stage(self, url: str, force: bool = False) -> bool:
    # Skip if same stage already loaded
    if not force and self.current_stage_url:
        if self._stage_path_key(url) == self._stage_path_key(self.current_stage_url):
            logger.info("Stage already loaded, skipping reload: %s", url)
            return True
    # ... proceed with actual load

The force=True parameter is used by _handle_reset_stage so explicit resets always work. Without this optimization, duplicate or reconnect-time openStageRequest messages can trigger redundant reloads.

Caveats
  • Use os.path.normcase(os.path.abspath(...)) for path comparison — not string equality.

  • Keep import os at module level, never inside _load_stage().

  • The frontend should not send a default openStageRequest on connect. Let the server's delayed push_initial_state send the current openStageResult and root children after the data channel opens.

  • The frontend should send openStageRequest only for explicit scene switches, file opens, resets, or reloads.

  • A same-stage openStageRequest should fast-return success and current root state without calling reset_stage() or open_usd().

Do Not Block The Render Loop

Loading can be slow because USD composition, texture/material discovery, and shader compilation may continue after open_usd* starts. In streaming apps, do not run the full load synchronously on the WebRTC message handler or block frame production for more than a few seconds.

Use a background loading thread plus a stage_lock around renderer mutation. While the lock is held for open_usd*, add_usd_reference*, remove_usd(), or reset_stage(), the render loop should skip renderer.step() and keep streaming the last good frame.

Rules

  • clippingRange is float2 clippingRange = (near, far), not separate .near/.far attributes.

  • Use open_usd() for file-backed root stages and open_usd_from_string() for generated root USDA. Both replace the active root.

  • Use add_usd_reference* only for additive content under a unique prefix_path; keep the returned handle if you will remove it later.

  • Camera path must match the path used by camera-controls when writing omni:xform; the reference streaming camera is /OVCamera.

  • The user scene is never modified by wrapper/session loading.

  • Write composite files in the same directory as the user USD so relative textures, MDL files, and sublayers resolve correctly.

  • Keep composite files alive for the server lifetime; ovrtx can perform async texture/material loading after open_usd() returns.

  • If the user stage has a camera, copy focal length, horizontal/vertical aperture, clipping range, and transform for visual consistency.

  • Use PrimMode.CREATE_NEW for camera omni:xform; the camera may only have authored xformOp:*.

  • Add InstanceSegmentationSD only when a debug/display segmentation AOV is needed. 0.3 picking uses ovrtx pick queries and resolved path IDs; it does not require a token-map RenderVar.

Failure Modes

  • Unable to find RenderProduct prim: wrapper/session render pipeline missing or wrong render product path.

  • Black frame: camera path invalid, resolution missing, or RenderVar sourceName wrong.

  • Broken textures after wrapping: composite is in /tmp instead of beside the asset.

  • Textures fail later after initial load: composite was deleted too early.

See also: ovrtx-rendering, render-settings, camera-controls, stage-management, streaming-server.

Requirements and compatibility

Written for: OpenAI Codex

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.

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
OpenAI (Official)
Distributed by
OpenAI
Source
openai/plugins · plugins/nvidia/skills/omniverse-realtime-viewer · commit d420d13
Licence
Apache License 2.0 (redistribution allowed)
  • LICENSE (written by AutoGineer)
  • ATTRIBUTION.md (written by AutoGineer)
Synced

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

Download ZIP

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