Govie - 3DIT's Viewer Technology

Govie® - The core of all applications


Govie is a 3D viewer technology for 3D, CAD, BIM, and geospatial data in the browser, consolidating 3DIT's expertise.

It includes a WebViewer for rendering, an asset pipeline for processing, and an integration setup for existing web applications.

The viewer runs natively in the browser without any installation or plugin, and can be embedded into any frontend ecosystem. Compatible with desktop, mobile, VR, and AR.

Visual QualityInteractivityPerformanceIntegrationSecurityCompatibility

These companies rely on the Govie 3D-Viewer Technology for their B2B communications

Display machinery, complex animations, and particle flows photorealistically on the web!

Visual Quality

Govie is a 3D viewer technology for complex visualizations of 3D, CAD, BIM, and geospatial data. It runs natively in the browser without installation and can be embedded in any frontend ecosystem – desktop, mobile, VR, and AR. The engine combines photorealistic rendering techniques with technical visualizations: PBR material systems for physically correct surface representation, advanced lighting algorithms, and Non-Photorealistic Rendering (NPR) for hidden-line rendering, cross-sections, and construction drawings.
Photorealistic 3D viewer in the browser showing an Egger iris diaphragm control valve in stainless steel
PBR materials reproduce stainless steel, coatings and glass photorealistically in the browser.
3D visualisation of a control cabinet in the rain with water droplets on the metal surfaces
Surface detail and particle effects such as rain and water film are generated in real time.
3D city scene with green roofs in the rain, including weather forecast and water balance in the viewer
Environment lighting and weather effects determine the scene's shadows and mood.

Technical Competencies

  • Physically Based Rendering (PBR) & Material System: Microfacet-based shading with albedo, roughness, metalness, normal maps, and parallax occlusion mapping. Material features: Clearcoat (multi-layer coatings), Sheen (silky-looking textiles), Iridescence (color shifting). Real-time switching between PBR materials and functional modes (X-ray, heatmaps, cross-sections) without scene reloading.
  • Image Based Lighting (IBL) & Environment Mapping: HDRI textures with precomputed Radiance Maps and Irradiance Probes for global illumination. Specular reflections and indirect shadows without real-time light sources.
  • Shadowing & Depth Effects: Shadow mapping with Percentage Closer Filtering (PCF) and Cascade Shadow Maps. Local shadowing via SSAO. Optional SSR and SSGI for specular and diffuse spatial effects.
Display complex information interactively across different scenarios!

Interactivity

The viewer treats the 3D scene as an interactive information layer. Selected objects can be transformed directly in space. Predefined scenarios switch layers, animations, and states via a state manager without reloading the scene. Additionally, any external data layers can be spatially embedded: annotations, hotspots, sensor values, or heatmaps are overlaid as 2D screen elements on top of geometry and remain correctly positioned during camera movement. For immersive use cases, a complete WebXR integration is available.
3D model of a process plant with overlaid sensor readings for pressure and temperature
Screen-space labels bind measured values to objects and stay accurately positioned as the camera moves.
Exploded view of a centrifugal pump with an interactive parts list in the 3D viewer
Object selection via picking hides assemblies and opens the exploded view.
Interactive 3D scene: users switch between roof structures and compare the water balance
Predefined scenarios switch variants and layers without reloading the scene.

Technical Competencies

  • Object Manipulation & Transform Controls: Objects can be moved, rotated, and scaled directly in the scene via gizmos and axis-bound handles. Transformations occur in local or global coordinate spaces, with support for snap-to-grid and free input constraints.
  • GPU-side Picking & Raycasting: Object selection via GPU-based picking (framebuffer readback) or CPU-side ray-triangle intersection. Hit accuracy at mesh level, even with complex overlaps and thousands of objects. Multi-select via box selection possible.
  • Screen-Space UI & Labeling: Annotations, labels, and hotspots are spatially bound to objects and remain correctly positioned during camera movement (Css2D renderer). Occluded labels are automatically hidden or sorted by priority. Supports text, images, and HTML-based UI elements.
  • State Management & Scenario Control: Predefined scenarios control visibility, animations, and material states without reloading the scene. Event-based triggers enable contextual reactions to user interactions (selection, hover, click). WebXR integration for immersive controls (hand tracking, controller input) in VR and AR applications.
Fluid rendering of large volumes of 3D data directly in the browser!

Performance

Large scenes with heterogeneous data sources are rendered efficiently on end devices through two complementary strategies: An offline pipeline optimizes geometries before delivery through reduction and level-of-detail generation. Govie's runtime engine intelligently selects visible detail levels, discards geometry outside the viewport, and loads missing data on demand. This keeps memory consumption constant, even with complex data sources (CAD, BIM, LIDAR, GIS) on mobile and desktop devices without user-side optimizations.
3D terrain model with overlaid layers for railway, roads, vegetation and city model
Streaming and culling keep large geospatial datasets with multiple layers running smoothly.
3D scene with rain effects computed in real time on a control cabinet model
Particle and weather effects run without dropping the frame rate.
Large 3D plant scene of automated truck loading and unloading built from reduced CAD models
Levels of detail and instancing render thousands of repeating parts at once.

Technical Competencies

Offline Pipeline

  • CAD Reduction & Defeaturing: A pipeline specifically designed for CAD data optimizes the scope of BREP models. Defeaturing removes irrelevant construction details, proxy geometries replace complex structures, shrinkwraps (outer shell approximation) drastically reduce polygon count without losing visual details.
  • LOD Generation & Texture Baking: Multiple discrete detail levels are automatically generated from the source mesh. Surface details (unevenness, scratches, textures) are transferred via texture baking into normal and roughness maps, so low-poly meshes appear visually detailed.
  • Spatial Hierarchization & Clustering: Regardless of source format, all objects are organized into a unified spatial hierarchy (octree, quadtree, 3D Tiles per Cesium standard). This hierarchy enables efficient culling and streaming at runtime – only visible or nearby tiles are loaded.

Online Rendering Techniques (Runtime)

  • LOD Selection & Distance Culling: The renderer automatically selects the appropriate detail level per object based on camera distance and available GPU memory. Objects beyond viewing distance are completely skipped (frustum culling). Continuous LOD transitions prevent pop-in artifacts.
  • Visibility Culling & Occlusion: Objects outside the view area or behind other geometries are skipped. Non-visible content generates no GPU load. Hierarchy-based culling enables processing of large scenes (>1 million objects).
  • GPU Instancing & Batch Rendering: Similar objects (screws, supports, repeating elements) are rendered as single GPU draw calls (instancing), regardless of count. Multiple meshes are combined into large batches to minimize render overhead.
  • Streaming & Adaptive Compression: Scene tiles are reloaded on demand. Visible tiles are transmitted with priority, missing data asynchronously in background. Mesh-based compression (Draco, Meshopt) reduces file size by 90% without losing geometry quality.
Seamlessly and flexibly integrate the 3D viewer into existing web apps

Integration

The Govie 3D engine is integrated as an encapsulated module via an API bridge into existing web applications. The viewer runs isolated in the browser and communicates bidirectionally with the host frontend (e.g., React, Vue, Angular) – for camera control, object selection, and dynamic data updates without scene reloading. Automated asset pipelines prepare CAD data before delivery, while standardized WebSocket channels synchronize real-time data streams (IoT sensors, metadata) spatially with the 3D scene.
Grafana dashboard with an embedded 3D model of a heat pump and live measurement values
The API bridge connects the viewer to Grafana and feeds measured values into the scene.
Model overview with automatically prepared CAD variants of valves and actuators
Automated pipelines convert CAD variants and publish them to the viewer.
Yacht configurator on a website: coating colours for hull and superstructure are selected on the 3D model
The host frontend controls the embedded viewer and switches coating variants.

Technical Competencies

  • Web Embedding & Frontend Integration: The viewer runs as an encapsulated module (Web Component) in the browser and communicates via bidirectional API bridge with the host frontend (React, Vue, Angular). Frame-synchronous communication enables synchronization of camera position, object selections, and attribute data between viewer and application.
  • Integration of various CAD formats: Govie supports proprietary and standardized data formats (STEP, IGES, Parasolid, Catia, Solidworks, Inventor, FreeCAD). All formats are converted into a unified runtime format and automatically optimized for web display (geometry reduction, material normalization, texture adaptation).
  • Real-time sync & live data binding: Standardized WebSocket channels synchronize real-time data streams (IoT sensors, measurement data, states) spatially with the 3D scene. Data synchronization (e.g., of position and viewing direction, user comments, etc.) also enables multi-user collaboration.
Schema of the Govie embed: an iFrame in the website DOM communicates with the host page via postMessage
An iFrame embeds the viewer, postMessage carries commands and events.
Protection of CAD data and intellectual property

Security

All geometry data displayed on modern graphics hardware is fundamentally readable. Primary protection measures are therefore applied before transmission: CAD construction data is destructively simplified in the pipeline so that, for example, only the outer shell is transmitted. Precise dimensions, construction history, and features are thus not reconstructible. For particularly sensitive 3D data, alternative representations are available that no longer contain usable construction data.
Schema of the CAD protection pipeline: hull model, simplification, Gaussian Splatting and pixel streaming
The pipeline strips design data and ships only hull models, splats or pixel streams.

Technical Competencies

  • IP Protection & Geometry Obfuscation: Feature removal removes feature trees, parameter values, and dimension chains. Geometric obfuscation (added micro-deviations) makes exported meshes unusable for reverse engineering.
  • Alternative Representations: Gaussian Splatting or point cloud representation instead of polygon geometry. These representation forms prevent reconstruction of precise CAD geometries while still offering complete visual quality.
  • Pixel streaming: With this type of solution, 3D content is rendered server-side and transmitted to the client only as a compressed video stream. Geometry and raw data remain on the server.
Platform-independent compatibility for BIM, GIS, and LIDAR data as well as proprietary CAD data

Compatibility

The Govie 3D engine processes heterogeneous data sources through two complementary strategies: A conversion pipeline transforms proprietary and standardized formats (CAD, BIM, GIS, LIDAR) into unified internal representations (meshes, point clouds, 3D Tiles).

A spatial coordinate transformation unites georeferenced data from various source systems into a global coordinate system. The viewer displays all data sources geometrically and spatially correct, regardless of format and origin.
Measuring roof areas on a 3D model of a church built from drone imagery
Photogrammetry data is converted and can be measured directly in the viewer.
The same scene compared as IFC, DTM, LiDAR, drone imagery and point cloud
The conversion pipeline merges IFC, DTM, LiDAR and point clouds into one scene.
3D plant site with labelled buildings and spatially located drone imagery
Coordinate transformation places models georeferenced in their real surroundings.

Technical Competencies

  • Data Conversion Pipeline: Proprietary and standardized formats (STEP, IFC, DWG, RVT, LAS) are automatically converted into a unified viewer format. Normalization of orientation, dimension, and geometry types enables uniform display of all sources.
  • Geodetic Coordinate System: All input data is transformed into a shared ECEF (Earth-Centered, Earth-Fixed) coordinate system, allowing georeferenced data from different origins to be displayed with geometrically correct overlay.

Everything you need to know about interactive 3D product presentations

Frequently asked questions

What do you need from us to create an interactive 3D product presentation?
How will our know-how be protected in an interactive 3D product presentation?
We have hundreds of product variants. Won't 3D product presentations be too expensive?
What if we only have a very small budget?
We are familiar with 3D. Can we reduce costs by doing some work ourselves?
Our agency also does 3D. What's the difference?
And what about VR, AR, MR, XR and the Metaverse?
Get your free demo

Get started with your interactive 3D product presentation!

Contact us today for a personal introduction

We've noticed you've set a different preferred language. Do you want to switch to the English site?

Switch to the English site

Or select your preferred language from the list below:

DeutschEnglish