2026 has been fairly full on so far, in a good way I should add. Lots of CityEngine work across a mix of knowledge transfer sessions and a couple of fairly large client projects. I’ve also been trying to get my head around some of the newer workflows Esri keep pushing into the product. As always, the flexibility of CityEngine remains the key selling point. Whether it is GIS into CityEngine and out to Scene Viewer or pushing things through to Twinmotion or Unreal, there’s still nothing quite like it for rapid urban scenario iteration.
Anyway, this post is about CityEngine 2025.1, which dropped back in December 2025. I know 2026.0 is just around the corner as we’ve got it out for a beta release currently (no I can’t speak about it yet officially!). Anyway, I have to say I had similar feelings to previous releases. Calling this a .1 feels a bit misleading given some of the things tucked inside, but perhaps that is just me wanting neat versioning rules that the real world ignores. The usual disclaimer applies. These are just my thoughts based on how I use the software. Yes I am an Esri partner, no I am not paid to write this, and yes I will moan about things I still want improving.
CGA Geometry Changes
Right, straight to the big one. This is probably the most important update in this release in my opinion. We now have proper geometry modification inside CGA, mainly through things like the new modify operation and related changes. In plain terms, this means we are no longer quite as boxed in by traditional split based workflows. You can now tweak parts of geometry and then stitch it all back together cleanly.
I have to admit I have mixed feelings here. On one hand this is brilliant. It opens up much more expressive architecture, especially anything that is not clean and orthogonal. On the other hand, it does feel like CGA is creeping closer to a modelling tool rather than staying purely procedural. That said, for real world projects where clients want “interesting” buildings, this is a massive win.
Visual CGA and Facades
This continues to evolve and it is quite clear where Esri are heading. The Visual CGA Editor now has facade components and a much richer component system. You can build up floors, windows, patterns and all that good stuff without writing much, if any, code.
Now, I am still very much a code first person when it comes to CityEngine as are most constant users of it. I like seeing exactly what is going on. But I can absolutely see the value here, especially for teams where not everyone wants to touch CGA.
It does make me wonder how this plays with more advanced rule packages though. I suspect there will be a bit of friction between Visual CGA users and those of us who like everything parameterised and scripted! To be honest I would have liked to see improvements elsewhere in terms of workflows and file support (stares at FGDB imports not working for arcs and circles).
Python 3 (finally… sort of)
This one will get attention from anyone who has been stuck in Jython land for years. There is now a Python 3 API in beta. I think we all knew this was coming eventually. Jython has been hanging around for a long time and it has always felt like a limitation, especially when trying to integrate with modern Python tools. A few thoughts:
- It is about time
- It being beta means I am not touching it for client work just yet
- Long term this is a very good thing
- I hope we can still interact with the Eclipse IDE a bit (jython allows access to java components)
I am particularly interested in how this will open up integration with the wider Python ecosystem. There is a lot of potential here for automation, data processing and linking into other tools in a way that has always been a bit painful.
Street Designer Improvements
Street Designer got its big debut in 2025.0, so this release is more about refinement. Some of the more useful bits:
- Better access to lane attributes for rule authors
- Improvements to lane editing and visual clarity
- General performance improvements when editing streets
- Nice quality of life tweaks in the inspector
You can now get more structured data out of lanes as well, which is handy if you are doing anything data driven downstream. I am still warming to Street Designer, I can see where it is going and it is very powerful. I’ve also made a script tool to take street cross section designs from StreetScape Studio into CityEngine, making the creation of streets not just a CityEngine user’s challenge/task.
General Modelling Workflow
There are also improvements around massing and the jump from simple shapes into more detailed buildings. Esri have included new example content that shows off more contemporary architecture styles and how the new CGA tools can be used. This is good to see. One of the common criticisms of CityEngine has always been that everything looks a bit samey out of the box. These newer workflows definitely help push things in a more design focused direction.
A Quick Note on Direction
This release feels like part of a bigger shift rather than a standalone update. A few themes stand out:
- Less reliance on strict CGA coding
- More visual and component based workflows
- Moving toward modern Python
- Richer geometry handling
It is all positive, but it does raise questions about where the balance ends up between procedural control and ease of use.
Summary
So, is this a big release? Honestly, yes even if it is labelled as a ‘.1’. The highlights for me:
- New CGA geometry capabilities are a genuine step forward
- Visual CGA is becoming a serious tool
- Python 3 is finally on the horizon
- Street Designer continues to improve
Should you upgrade?
Yes, with the usual caveats. I would be cautious using the Python 3 side of things in production yet, but the CGA improvements alone make this worth a look. As always, credit to the CityEngine team. There is clearly a lot of work going on behind the scenes. Now if we could just get a bit more love for scripting workflows and maybe a clearer long term plan around Python, that would make me very happy indeed.
