Registration is open for the 2026 Cesium Developer Conference! Save your spot.

Cesium for Omniverse v0.28.0

Cesium for Unity Samples v1.23.0

CesiumJS 1.139.1

CesiumJS

Matom.AI is applying AI to city‑scale point clouds, extracting meaningful insights from complex geospatial data for city management, power line inspections, and forest analysis. A Cesium Certified Developer, Matom.AI uses CesiumJS to build digital twins from massive datasets, including custom terrain, CityJSON, photogrammetry, orthophotos, and point clouds. The real innovation of Matom.AI’s platform is their AI‑driven approach to analyzing this data.

Matom.AI uses AI tools to segment point clouds and identify common elements such as trees and lamp posts. Here, they’ve automatically classified and modeled urban assets in a digital twin of Kaunas, Lithuania. Courtesy Matom.AI. 

Because CesiumJS and 3D Tiles make massive geospatial datasets easy to stream and visualize, Matom.AI can apply AI directly to raw city data, eliminating the need for labor-intensive modeling and enabling scalable, continuous analysis.

While a traditional digital twin requires developers to manually clean data, hand-build models, simplify, and then visualize, Matom.AI has flipped this order. They bring point clouds directly into CesiumJS and visualize first, then analyze at scale and structure with AI. These digital twins make it possible for teams in different locations to work with complex geospatial data directly in the browser, where they visualize large‑scale 3D datasets in a unified, high‑performance environment.

By fusing datasets captured at different times and from different perspectives, Matom.AI’s digital twins provide a more complete view. Cesium enables these diverse data sources to be visualized seamlessly within a single platform. Streamed in the browser as 3D Tiles, the data supports shared workflows, custom analysis, and reporting. The platform also supports real-time data collection through a mobile app, allowing field teams to capture information and add updates directly to the digital twin.

“CesiumJS allows us to deliver a truly collaborative digital twin that all stakeholders can access simultaneously. Everything streams directly in the browser, so teams across a city can make decisions together in real time.”

– Julius Gelšvartas, Chief R&D Officer and Co-Founder at Matom.AI

Matom.AI’s platform integrates AI-driven point cloud classification and instance segmentation, automatically detecting benches, poles, traffic signs, trash cans, and other urban assets. Each object can be analyzed individually and converted into structured models, making visualization and inspection more efficient. Built-in tools also calculate core characteristics of each element, such as location, height, and width, simplifying inventory management and infrastructure assessments.

AI-powered image segmentation, combined with measurement analysis tools, provide quick insight into large datasets, as in this detail view of a bus stop in Kaunas, Lithuania. Courtesy Matom.AI.

One of the most valuable capabilities Matom.AI provides is AI-powered change detection. By comparing scans captured at different points in time, the system highlights changes in buildings, vegetation, and city assets, helping users identify issues earlier and prioritize maintenance before small problems grow into larger challenges.

Here, Matom.AI highlights discrepancies between two datasets: one captured in winter and the other in summer in Kaunas, Lithuania. Courtesy Matom.AI. 

By making point clouds easier to analyze and act on, Matom.AI enables cities to collect data more frequently and maintain a continuously up‑to‑date digital twin. To visualize and analyze your point clouds in a geospatial environment and easily share with colleagues, bring them into CesiumJS using 3D Tiles.

If you are building creative solutions with Cesium, consider applying for the Cesium Certified Developer program, an opportunity to demonstrate your expertise and connect with other developers in the community.

Matom.AI Extracting AI Insights from Point Clouds in CesiumJS

In December 2025, we shared our progress on AI-driven semantic search for Cesium Sandcastle. This feature is now generally available. 

Sandcastle has long been a cornerstone of the CesiumJS developer community: showcasing new features, sharing ideas, learning, prototyping, experimenting, and validating code. Its accessibility means developers can quickly find, build, and share code examples relevant to their projects, with demonstrations on how to implement features in the CesiumJS library. 

With the semantic search, users are now able to make a wider range of searches to more efficiently find sandcastles relevant to their queries. You no longer need to exactly match the name, description, or code to find applicable examples. 

Examples of semantic search in Sandcastle.

More efficiently find sandcastles relevant to your queries with semantic search.

Semantic Search in Cesium Sandcastle

This is powered by an open-source embedding model, which embeds the user's query and compares it to the semantic embeddings that are now precomputed for the full code contents for all sandcastles. This open-source embedding model is run locally on the end user's hardware, so your search data never leaves your device, just as before. By leveraging cutting-edge models optimized for small size, search performance remains fast. 

The semantic search functions as a hybrid with the previous search, so any existing search results will still appear as before. To opt out of the semantic search, you can disable the "Semantic Search" toggle in settings. 

Try the new semantic search here, and share feedback on GitHub. 

AI-Driven Semantic Search in Cesium Sandcastle Now Available

Last month Camptocamp released Geo Challenge, a browser-based game built on CesiumJS that challenges users to identify randomly selected locations in Switzerland, France, or Germany.

Geo Challenge awards players points based on how close their guesses are to a 3D view’s actual location. Courtesy Camptocamp.

The game uses high precision datasets—swisstopo’s high‑resolution topographic data for Switzerland and Google Photorealistic 3D Tiles served by Cesium ion for France and Germany. The game is purely for fun, but the simple interface belies the sophistication of the massive datasets, 3D tiling pipelines, and geospatial rendering that make it possible. By using CesiumJS, Camptocamp transforms complex 3D data engineering into a playful, interactive experience that invites users to feel the value of high-precision geospatial data.

GeoChallenge by Camp to Camp

“With Cesium we’re able to give players a view of their countries with such accurate terrain that many are able to recognize the spot instantly. Players aren’t just guessing coordinates; they’re exploring.” 

- Frédéric Junod, Camptocamp developer and the game’s creator

In Switzerland, Geo Challenge uses swisstopo’s high-resolution datasets, rendering every peak and valley of the country. Courtesy Camptocamp.

Camptocamp shares Cesium’s commitment to open, collaborative technology and has extensive experience building with CesiumJS. A decade ago they worked with Switzerland’s Federal Office of Topography, swisstopo, to build an interactive 3D map of the entire country using CesiumJS. They were also an early adopter of Cesium’s subsurface rendering capabilities, used to show geological features below ground. Camptocamp is a Cesium Certified Developer organization, and they continue to provide geospatial services for companies and governments in Europe.

The Swiss players in particular are intensely competitive (and very good!), with many players having near perfect scores. Courtesy Camptocamp.

To see what experiences you can create with 3D geospatial data, try out CesiumJS.

Cesium Powers Camptocamp’s Geo Challenge Game

We’re excited to add support for panoramic imagery in CesiumJS, enabling developers to add high-resolution 360° panoramic imagery from sources like Google Maps into their 3D geospatial applications. Viewed in a 3D scene and combined with other 3D assets, panoramic imagery provides immediate geospatial context, making the ground‑level scene easier to understand, navigate, and relate to surrounding terrain and assets.  

This CesiumJS sandcastle shows how to include a panoramic image in a CesiumJS app.

Google Street View contains a vast global resource of panoramic imagery. We created a helper class to load panoramic Google Street View images from anywhere in the world directly into a CesiumJS scene. This helper class is used in the sandcastle example below. Google Maps also makes available a Street View availability layer, which is included in the example to indicate where Street View imagery is available for loading. 

This CesiumJS sandcastle shows how to include a panoramic image from Google Street View in a CesiumJS app.

CesiumJS supports the display of panoramic imagery in equirectangular or cube map formats. Both can be geolocated and oriented within a scene to align with other geospatially accurate features and integrate cleanly with all 3D rendering capabilities in CesiumJS. 

A panoramic image in equirectangluar format and cube map format. Photo by David Iliff. License: https://creativecommons.org/licenses/by-sa/3.0/.

CesiumJS supports bringing your own panoramic imagery to render within a scene, as well as loading panoramas from external providers such as Mapillary. In the example below, panoramic images were captured alongside LiDAR point cloud data using a mobile mapping system, and CesiumJS allows both datasets to be visualized together in the same 3D scene. 

A CesiumJS scene featuring a panorama image aligned with a LiDAR point cloud, courtesy of Bentley iTwin Capture. 

Split Screen: Panoramic and Point Cloud

Ready to enhance your 3D scenes with high-resolution, geospatially accurate panoramic imagery? Check out the tutorials on displaying panoramic images and using Google Street View to get started. 

Introducing Panoramic Imagery in CesiumJS 

In February, we attended Geo Week in Denver and discussed interoperability and open standards for Gaussian splats in the RTC AECO Spotlight. This week we’re proud to host the 134th OGC Member Meeting—let us know if you’re in Philadelphia! Plus: two more weeks of early bird devcon registration. 

We made improvements across Cesium this month. To get notifications about our monthly releases, you can subscribe to the Cesium release roundup thread on our community forum.

CesiumJS 1.139 is now available. Highlights of the release include:

Added panorama support via new EquirectangularPanorama and CubeMapPanorama classes, along with GoogleStreetViewCubeMapPanoramaProvider for loading cube map faces from the Google Street View Static API and rendering them in a cube map panorama. #13153 

Added more depth testing options for billboards and labels. #12994 

Added support for more metadata types via property textures in custom shaders. #13135 

Modified Sandcastle application to use a hybrid text and semantic, embedding based search. #13090 

Fixed Gaussian splat race conditions in snapshot/sort updates. #13016 #12965 

Fixed flashing when rendering multiple Gaussian splat primitives. #12967 

Part of a 360-degree panorama image of the interior of Laon Cathedral, France, visualized in CesiumJS. Photo by David Iliff, licensed under CC BY-SA 3.0.

Cesium for Unreal v2.23.0 is now available. Highlights of this release include:

Added support for loading tilesets with the KHR_gaussian_splatting extension. 

Added support for styling with tileset statistics on UCesiumFeaturesMetadataComponent. 

Deprecated AddCamera, RemoveCamera, UpdateCamera, and GetCameras on ACesiumCameraManager. Instead, use the new AdditionalCameras array to directly add or remove cameras. 

Added UsePlayerCameras, UseEditorCameras, and UseSceneCapturesInLevel to ACesiumCameraManager to toggle the automatic detection of cameras in the level for tile selection. 

Added ACesiumCameraManager::SceneCaptures for more explicit control over which ASceneCapture2Ds are used for tile selection. 

Added DetectTransformChanges to UCesiumGlobeAnchorComponent to toggle whether the globe anchor should update itself after changes in its Actor's transform. 

Build now uses the tinyxml2 library from Unreal's SDK on Linux and Windows if it is available. This eliminates a source of conflicts with other plugins that also use tinyxml2. 

Removed Cesium3DTileset::getFeaturesMetadataDescription. This can be directly retrieved from its attached UCesiumFeaturesMetadataComponent instead. 

In addition to the above, this release updates cesium-native from v0.57.0 to v0.58.0. See the changelog for a complete list of changes in cesium-native. 

Check the Cesium for Unreal release and the Cesium for Unreal Samples release for the full list of updates.

Download the Cesium for Unreal plugin from Epic Marketplace

Hines Rowing Center in Conshohocken, PA, visualized in Cesium for Unreal using Gaussian splats produced by iTwin Reality Capture.

Cesium for Unity v1.23.0 is now available. Highlights of this release include:

Added CesiumGeoJsonDocumentRasterOverlay, allowing stylized GeoJSON to be rasterized and draped over terrain and other 3D Tiles. 

Added CesiumGeoJsonDocument to parse and manipulate GeoJSON documents in code. 

Added CesiumGeoJsonObject to access individual features, geometries, and collections within a GeoJSON document. 

Added CesiumVectorStyle, CesiumVectorLineStyle, CesiumVectorPolygonStyle, and CesiumVectorPolygonFillStyle structs for styling vector data. 

Check the Cesium for Unity changelog for the full list of updates, and follow our Cesium for Unity tutorials to get started. 

A footprint of a proposed design model projected onto Cesium World Terrain using CesiumGeoJsonDocumentRasterOverlay.

Cesium for Omniverse v0.27.0 is now available. Highlights of this release include:

Cesium for Omniverse now requires Kit 107.3 or above (Isaac Sim 5.1.0 or above). 

Peruse the changelog for the full list of updates, and get started with our Cesium for Omniverse tutorials. 

Download the Cesium for Omniverse extension

Download the Cesium for Omniverse samples

**Cesium ion SaaS**
- Coming this week: additional options for running photo reconstruction jobs. These will make it possible to tune your automated reality modeling pipelines to better meet your requirements. 

**Cesium ion Self-Hosted**
- Coming this week: new version of Cesium ion Self-Hosted that includes recent updates to the Cesium design and reality tilers. 

**3D Tiling Pipeline** 
- The Design and Reality Tilers are now packaged with the tilers and tools in the 3D Tiling Pipeline. Versions have been unified, and all tilers are now on v5.0.1. 
- Added `terrain-and-imagery-tiler`, which produces 3D Tiles 1.1 output. 
- Removed `quantizedmeshto3dtilesconverter`; it has been replaced by `terrain-and-imagery-tiler`. 
- Fixed an issue where invalid attributes could be generated when converting from OBJ to glTF.

    - **Design Tiler** 
      - Improved rendering of IFC elements by giving sub-assembly elements the bounding box of the full assembly. 
      - Added more graceful handling of IFC lines that don’t meet the IFC specification. 
      - Added support for parsing eastings, northings, and orthogonal height from IFC files. 
      - Fixed generation of IFC swept circular geometry for better shading and performance. 
      - Fixed issues processing B-splines in IFC files. 
      - Fixed issues processing some geometries with swept circles. 

    - **Reality Tiler** 

      - Fixed an issue where duplicate vertices in input models could lead to inefficient textures and larger tilesets than necessary. 
      - Fixed a regression where duplicate vertices in the input would be written to tiles, rather than being filtered out. 
      - Significantly reduced memory usage for models with spatially incoherent textures. 

    - **Buildings Tiler**
      - Fixed an issue where some images in a glTF file had an invalid format and/or incorrect mime type. 

Trailamo uses CesiumJS to bring open geospatial data to mountaineers so they can plan safer, more informed routes anywhere in the world with advanced 3D terrain analysis and layered datasets. 

BADASS Studios uses Cesium for Unreal to create a premium digital fan experience for the E1 Series. 

Creating an XR Connection to the E1 Series with Cesium for Unreal

Matteo Massimi is an avid mountaineer. So when he took his most recent trip to Central Asia, there was nothing unusual about the hikes. What was new was his ability to plan the expedition in the app he created with Cesium: Trailamo.

Trailamo uses CesiumJS to bring open geospatial data to mountaineers so they can plan safer, more informed routes anywhere in the world with advanced 3D terrain analysis and layered datasets.

Trailamo demo

Trailamo is built on CesiumJS and tailored to the way an explorer like Matteo, looking for off-the-map adventures, plans outdoor experiences. With a highly precise WGS84 ellipsoid and the ability to render high-quality geospatial datasets, CesiumJS gives Trailamo a solid foundation for its sophisticated terrain analysis and numerous planning tools. 

Explorers can create their own routes in Trailamo, dropping waypoints and comparing against public routes, as in this climb of the southwest face of the Matterhorn. Courtesy Trailamo.

“I’m not a software developer by training, but I knew what features I needed in the ultimate mountaineering app. So I looked for what tools were available to a noob like me, and with Cesium I could build exactly what I needed.”  

Trailamo offers global terrain and imagery for planning an outing anywhere on the planet. Beginning with Cesium World Terrain, users can then choose from a variety of imagery layers, such as topographic maps or satellite imagery. For a photorealistic view of well-known areas, they can also view global Google Photorealistic 3D Tiles.

Trailamo aggregates open data, using CesiumJS to give viewers access to public datasets for planning outdoor experiences, as in this view of Mont Blanc in the Alps from the French approach. Courtesy Trailamo.

Above this base, users can add overlays, such as live snow coverage from NASA GIBS, weather data from OpenWeather, ski pistes from OpenSkiMap, hiking trails provided by various collections of public GPX tracks, and more.

By combining terrain, imagery, and overlays, users can better understand upcoming adventures. Here hiking trails, publicly available routes, ski runs, and terrain contours are all shown with OpenTopoMap imagery draped over Cesium World Terrain at Fizkul Peak in Kazakhstan. Courtesy Trailamo.

Trailamo offers free accounts. Matteo, a strong supporter of open data, believes that high‑quality outdoor tools shouldn’t hide essential information behind steep subscription fees—especially when much of that data is generated by publicly funded government agencies as a public good. If citizens have already paid to collect it, he argues, they shouldn’t have to pay again to access it.

Terrain analysis tools, such as this calculation of avalanche angles at Aconcagua, Argentina, give adventurers a better understanding of their surroundings so they can mitigate their risk in unknown territory. Courtesy Trailamo.

Matteo has also written a number of tools for specialized terrain analysis. Because CesiumJS exposes the quantized mesh in its terrain pipeline, Matteo was able to calculate normal vectors for each triangulated surface and use that calculation to display altitude and slope gradient, avalanche angles, and terrain aspect and contours. 

Combining layers and overlays with tools can help users answer very specific questions about planned trips. In this screenshot of the Trans-Ili Alatau range in Kazakhstan, only peaks above a certain altitude are visible, and the slope gradient is colored to give a better understanding of the steepness of the terrain. Courtesy Trailamo.

Because Trailamo uses Cesium to aggregate numerous public datasets into a single app, users can combine them for better insights. This means that skiers can look at runs in their area overlayed with NASA‑collected snow coverage and check for avalanche angles in the backcountry for areas of risk. Hikers venturing into remote areas without widely published guidebooks and trails can look at the 3D terrain and overlay that with GPS tracks collected by OpenStreetMap to infer where others have traveled—quite literally following in past footsteps.

“Aggregating this open data with Cesium gives me the tools to have free outdoor experiences that would have been much more challenging, if not impossible, ten years ago. I’ve climbed new mountains because of Trailamo.” 

Matteo’s future plan is to continue to expand Trailamo’s data offerings, bringing in more open source data and allowing users to add their own open source imagery or overlays. 

To bring together the datasets you care about, sign up for a free Cesium ion account.

Planning Mountaineering Routes in CesiumJS with Trailamo

In January, we attended GeoBuiz in Denver and opened the call for presentations and early bird registration for the 2026 Cesium Developer Conference. Looking ahead, we’re thrilled to be a host of the 134th OGC Member meeting in Philadelphia in March. 

CesiumJS 1.138 is now available. Highlights of the release include:

Fixed jitter artifacts on Intel Arc GPUs. #12879 

Improved voxel memory usage by reworking Megatexture to use Texture3D. #12570 

Fixed multiple issues causing undefined pick results in 2D/CV scene modes. #13083 

Fixed label sizing for some fonts and characters. #9767 

Volumetric rendering with voxels in CesiumJS was optimized with WebGL2’s 3D Textures for reduced memory usage at runtime.

When adding a CesiumCartographicPolygon to the level via the Cesium panel, the polygon is now created in the current Editor view, rather than at the origin. 

Added ACesiumCartographicPolygon::SetPolygonPoints to shape a polygon spline from a list of positions in cartographic or Earth-Centered, Earth-Fixed coordinates. 

Re-enabled support for tileset water masks on macOS devices. 

In addition to the above, this release updates cesium-native from v0.56.0 to v0.57.0. See the changelog for a complete list of changes in cesium-native. 

Cesium for Unity v1.22.0 is now available. Highlights of this release include:

Fixed a bug that prevented CesiumGoogleMapTilesRasterOverlay from being added to a GameObject in more recent versions of Unity.  

Coming this week: The Design and Reality Tilers will now be packaged with the tilers and tools in the 3D Tiling Pipeline. Versions will be unified, and all tilers will be on v5.0.0. 

WorldLens uses Cesium for Unity to deliver high-performance, immersive VR experiences that stream global geospatial data to Meta Quest.

Blenheim Palace leveraged 3D Tiles created through the Cesium Design Tiler and Add-In for Autodesk Revit to create a digital twin in CesiumJS that has IoT integration to enable smart heritage conservation, predictive maintenance, and sustainability.

Blenheim Palace Performing Smart Conservation with Cesium

To modernize and synchronize operational-level planning, Sandtable uses CesiumJS, Cesium for Unity, and Cesium ion to combine traditional military planning documents and immersive geospatial context.

Over the last 20 years, technology and the nature of conflict have changed more rapidly than US defense policies and military procedures. To modernize and synchronize operational-level planning, Sandtable uses CesiumJS, Cesium for Unity, and Cesium ion to combine traditional military planning documents and immersive geospatial context.

For operational wargaming in the COA analysis step of the MDMP, as well as mission rehearsal, Sandtable Navigator provides a 4D environment built with Cesium for Unity. Pictured here are friendly and enemy units at the National Training Center at Fort Irwin, California, USA. Courtesy Sandtable.

For operational wargaming in the COA analysis step of the MDMP, as well as mission rehearsal, Sandtable Navigator provides a 4D environment built with Cesium for Unity. Pictured here are friendly and enemy units at the National Training Center at Fort Irwin, California. Courtesy Sandtable.

Sandtable: COA Analysis

Founded in 2024 by US Army veterans and researchers, Sandtable pairs AI with human judgment in its Mentat and Navigator prototypes to facilitate mission analysis, course-of-action (COA) development, and operational wargaming with more data, detail, and speed.

Planning, collaboration, and rehearsal tools tend to involve 2D imagery, spreadsheets, printed documents, a physical sand table, and the risks that come with putting a group of key people in the same place simultaneously. The company’s push to bring time-dynamic 3D to military planning results in interactive environments that fuse data types and sources to support planning and assessment for both human teams and autonomous systems, no matter where they are. 

During COA DEV, planners using Sandtable Mentat create, visualize, and share unit symbology, operational graphics, viewshed, and ranges of weapon systems in 4D. Courtesy Sandtable. 

“What we’re simulating and planning—like drones, counterdrone technology, and electromagnetic warfare—exists in 3D space in reality, and we need to be able to represent reality. That requires Cesium,” said Sandtable COO Brady Moore. 

And reality is complex: data comes from myriad sources, and conditions shift over time. With AI, the platform helps surface patterns, compare options, and stress-test assumptions, so commanders can focus their judgment where it matters most.

Planners drag and drop MIL STD 2525D symbols from predefined combat power data to build enemy SITEMPs and friendly COAs in Mentat. Courtesy Sandtable.

Planners drag and drop MIL STD 2525D symbols from predefined combat power data to build enemy SITEMPs and friendly courses of action in Mentat. Courtesy Sandtable.

Sandtable: Enemy SITEMPs and Friends COAs

Cesium is deployed across Sandtable’s solution to accelerate key phases of the military decision-making process (MDMP), particularly mission analysis, COA development, and rehearsal, with authoritative geospatial context and efficient data delivery:

Mentat, built with CesiumJS, incorporates written orders, concept sketches, and open datasets, e.g., for soil density and composition, on top of real-world terrain in a browser for rapidly adjusting COAs with AI.

Navigator uses Cesium for Unity for COA analysis and mission rehearsal in a shared virtual environment, wherever participants are, on a variety of devices. Instead of merely rehearsing plans produced in Mentat, Navigator feeds insights back into the planning environment—enabling continuous refinement rather than one-time execution.

Sandtable Mentat allows planners to determine which factors about the terrain are relevant to operations and visualize them together or separately. Courtesy Sandtable. 

With detailed geospatial data from Cesium ion and a range of public and government sources (for example, NOAA and USDA datasets), warfighters and autonomous vehicles know not only where hills and trees are but also microterrain features relevant to mobility and line of sight. This is important because factors like height of vegetation determine what a drone can—or can’t—see (depending on available data resolution and source quality). This data is primarily 3D Tiles, quantized mesh, and KML/KMZ, optimized, hosted, and streamed via Cesium ion for efficiency and interoperability. Cesium ion also enables Sandtable to be data agnostic, because 3D Tiles is an OGC community standard, and can use the Army's own Well Formed Format data when available. Sandtable uses Cesium ion SaaS while connected to the internet and anticipates leveraging Cesium ion Self-Hosted in future disconnected or denied environments, as well as secure enclaves.

Sandtable Mentat develops its own Modified Combined Obstacle Overlay (MCOO) using CesiumJS to visualize layers. Sandtable’s detailed datasets help determine terrain trafficability. Other terrain data includes canopy height, land use, soil attributes, slope, and elevation. Courtesy Sandtable.

Sandtable Mentat develops its own Modified Combined Obstacle Overlay (MCOO) using CesiumJS to visualize layers. Sandtable’s detailed datasets help determine terrain trafficability. Other terrain data include canopy height, land use, soil attributes, slope, and elevation. Courtesy Sandtable.

Sandtable: Modified Combined Obstacle Overlay (MCOO)

Critical to Sandtable’s work is embedding with the very people who will use the platform. During the Army’s xTechOverwatch competition in October 2025, the team tested Mentat on a tablet in the field with soldiers during the day and adjusted the capabilities and interface each night, returning the next day with an improved product. Sandtable developers were able to take feedback from concept to functional visualization and analysis in hours rather than days. This speed of development is enabled by adaptable, streaming-ready 3D Tiles and open-source runtimes. (The company’s CEO, Brian Hamilton, is a Cesium Certified Developer, and Sandtable was among the competition’s winners.)  

Watch Sandtable CTO Brent Lance’s 2025 Cesium Developer Conference talk to learn about the role Cesium plays across Sandtable’s solution for the MDMP. 

Military Planning Using Agentic AI and Simulations

Mentat and Navigator have been deemed Awardable on the Tradewinds Solutions Marketplace, making them eligible for government procurement and tightening collaboration between government and industry.

Use Cesium ion to securely fuse high-resolution imagery and terrain with unit-specific data for actionable information and collaboration. Get started with a community account.  

Author’s note: Sandtable’s COO, Brady Moore, is a former Cesium employee. 

Sandtable Modernizes and Synchronizes Military Planning with Cesium

Blenheim Palace, a UNESCO World Heritage Site and one of England’s most iconic estates, faces a monumental challenge: maintaining a 300-year-old building that welcomes over one million visitors annually, while also serving as a lived-in residence.

Blenheim Palace, a UNESCO World Heritage Site and one of England’s most iconic estates. Courtesy Blenheim Palace.

To keep up with these challenges, Blenheim Palace has invested in an IoT sensor network that monitors all sorts of data from the building, such as temperature, humidity, CO2 levels, and footfall.

The data, pouring in from over 30 platforms, became too much for their building management system (BMS) to take in and ended up siloed, often buried in folders. Maintenance teams, facing restoration costs of £4 million a year and overwhelmed by the sheer volume of data, found it nearly impossible to get a comprehensive picture of the state of the palace, let alone plan for the future.

At 300 years old, Blenheim Palace requires constant maintenance and reconstruction work. Courtesy Blenheim Palace.

That’s all changed now that they’ve developed a digital twin of Blenheim Palace, built with Cesium, that provides a single source of truth for the state of the palace. IoT sensors feed real-time alerts into a unified platform, allowing maintenance teams to visualize issues in context, plan restoration work proactively, and even receive automated notifications when environmental thresholds are breached.

Blenheim digital twin for smart conservation

Blenheim Palace’s digital twin was born with the announcement of the Cesium Design Tiler and Add-In for Autodesk Revit. The team at Blenheim saw that the 3D Tiles open standard would allow them to efficiently stream heavy Revit files into a lightweight environment, providing a frontend through which they could access all of the IoT data they were collecting.

Loaded in CesiumJS, the interactive palace model streams smoothly despite its size and complexity thanks to 3D tiling via the Cesium ion for Autodesk Revit Add-In.  Courtesy Blenheim Palace.

In partnership with the Heritage Innovation Laboratory Oxford and Oxford Brookes University, Blenheim’s team worked to make their BIM model the foundation of a digital twin. They sliced their architectural model into over 14,000 individual elements, each tied to metadata: material type, historical repairs, risk assessments, etc. Then, using the Revit add-in and the Design Tiler in Cesium ion, they tiled the model into 3D Tiles. Now with a model that retained all the detail and metadata of the original but tiled for smooth navigation, they loaded the 3D Tileset into CesiumJS.

The original palace model came from a scanned point cloud, corrected by architects and tied to floor plans. Courtesy Blenheim Palace.

With the Revit model as the frontend, the palace’s intricate structure became an interactive digital twin. The model of the palace is easy to navigate, accessible to both technical and non-technical staff. Users can filter views of the site, for example, by restoration needs deemed most urgent in a recent risk assessment.

Users can filter views of the site, for example, by restoration needs deemed most urgent in a recent risk assessment. Courtesy Blenheim Palace.

Behind the interactive model sits Blenheim’s entire IoT network. By centralizing data with Cesium and automating alerts, Blenheim Palace reduced manual oversight and enabled proactive maintenance, potentially lowering restoration costs by up to 10%.

Blenheim Palace’s digital twin has a growing database currently exceeding 150 million rows of environmental and building data, fed by IoT sensors integrated with BMS controllers. Courtesy Blenheim Palace.

Predictive maintenance guided by AI insights

The digital twin also contains archival data tied to the model. This historic view, developed by Clara Saliba, makes the model a portal to the archives, giving users access to decades of maintenance history and to machine learning models trained to identify patterns in stonefall, water ingress, and exfoliation. These machine learning models are expected to enable predictive maintenance and sustainable energy use.

“Heritage isn’t just about preserving the past—it’s about imagining its future. With Cesium ion, we turned our Revit model from a virtual replica into a digital twin of Blenheim Palace—a dynamic simulation that tracks condition and restoration data, identifies inefficiencies, predicts future needs, and transforms how we manage heritage through technology.” 

- David Green, Head of Innovation, Blenheim Palace

The digital twin isn’t just about capturing the current state of the building. The Blenheim team is using this data to reach specific goals: Blenheim Palace aims to be carbon net zero by 2027, and a precise, real-time view into the building allows them to track energy usage and optimize systems via smart controllers to meet this ambitious goal.

"The Digital Twin at Blenheim Palace has been the flagship project for the Heritage Innovations Laboratory Oxford since its inception in 2022. As soon as we began collaborating with Cesium in the winter of last year, we began making rapid progress in pursuing our collective goals of linking heritage science and digital twinning to arrive at predictive solutions for managing built heritage environments." 

- Mr. Sterling MacKinnon III, Heritage Innovations Laboratory Oxford

Future steps for the Blenheim Palace digital twin

Looking ahead, Blenheim team plans to add even more granularity to their digital twin: they’d like to slice it even further to capture more architectural details and to include household items like tapestries and furniture. They expect the model to eventually grow from its current 14,000 elements to over 50,000. In addition, they foresee being able to run frequent drone surveys of the building to provide regular updates on the state of the structure.

Blenheim Palace’s digital twin provides a model other heritage sites could use for smart conservation. Courtesy Blenheim Palace.

Blenheim’s team also built their digital twin with an eye to other heritage sites that could benefit from similar analysis. Few heritage sites around the world have the internal technical expertise that Blenheim has cultivated, and the Blenheim team has designed their digital twin to scale. They plan to develop their system into a broadly applicable platform other locations could also use, offering a model for sustainable digital conservation.

In Blenheim Palace’s digital twin, Cesium provides a spatial environment that has become the index to accessing, visualizing, and analyzing a host of data about the building. Here the palace is shown together with Cesium World Terrain and satellite imagery at its real world location. Courtesy Blenheim Palace.

To tile your own Revit models as 3D Tiles, making them streamable over the web or in runtime engines like Unreal Engine and Unity, download the add-in from the Autodesk App Store, the Cesium downloads page, or the GitHub repo.

In December, the team was in Orlando for I/ITSEC and Exton for our quarterly refactor. Then came all the jolly action with the Cesium-powered NORAD Tracks Santa app. 

Today, we opened the call for presentations for the 2026 Cesium Developer Conference. Submit your talk ideas, and we look forward to seeing you in June! 

CesiumJS 1.137 is now available. Highlights of the release include:

Fixes for positioning in workflows that delete and recreate labels. 

Fixes for texture coordinates in large billboard collections.

CesiumJS rendering a large collection of billboards with high precision, using data from the iTwin Feature Service.

Cesium for Unreal v2.22.0 is now available. Note that this release removes support for Unreal Engine 5.4; Unreal Engine 5.5 or later is required.

Other highlights of this release include:

Replaced the "Auto Fill" button on UCesiumFeaturesMetadataComponent with "Add Properties". This allows metadata properties and feature ID sets to be selectively added to the component, rather than adding them all in bulk. 

Added EnableDoubleSidedCollisions to ACesium3DTileset to allow intersections and collisions with geometry to also work when hitting "back" faces. 

Added an optional glTF model post-processing stage after the tile loading step to modify their models before they are displayed. 

Added support for refreshing Cesium ion login tokens after they expire and the refresh token is still valid. 

Fixed a material generation bug on UCesiumFeaturesMetadataComponent where user-made connections to If nodes were not preserved. 

Fixed deprecation warnings when compiling with Unreal Engine 5.5+ by replacing deprecated GetInputsView() calls with FExpressionInputIterator. 

This release also updates cesium-native from v0.55.0 to v0.56.0. See the changelog for a complete list of changes in cesium-native. 

A list of metadata properties available in the New York region of Cesium OSM Buildings.

Cesium for Unity v1.21.0 is now available. Highlights of this release include:

Setting CesiumGeoreference.ellipsoid at runtime will now correctly update the Georeference and associated transforms. 

Fixed a bug that could cause Cesium3DTileset::Update to be called more than once per frame in the Editor. 

Matom.AI Extracting AI Insights from Point Clouds in CesiumJS

AI-Driven Semantic Search in Cesium Sandcastle Now Available

Cesium Powers Camptocamp’s Geo Challenge Game

Introducing Panoramic Imagery in CesiumJS

Cesium Releases in March 2026

Planning Mountaineering Routes in CesiumJS with Trailamo

Cesium Releases in February 2026