AI is cooking us all.
To all,
I’ve read that SU’s programming code may be holding it back from certain UI requested updates & may have something to do with SU-Layout being slow to use. I don’t know anything too deep about programming code per-se, so I asked GOOGLE-GEMINI-AI to research this & compare SU & IngeTrazo in this regard. See below for the findings. Can experienced users who know about this subject comment on these findings, given AI can sometimes paint a distorted picture…
Code-Based Pros & Cons Breakdown
SketchUp (The Legacy Giant)
-
Underlying Architecture: Built on a decades-old C++ core framework paired with a Ruby API extension engine.
-
Code-Based Pros: Highly optimized core geometry processing engine refined over 25+ years, offering a stable and mature foundation for a massive commercial ecosystem.
-
Code-Based Cons: Entrenched technical debt makes deep multi-threading and modern UI overhauls difficult. Furthermore, LayOut suffers from sluggish performance because it operates as a decoupled companion app communicating across a heavy inter-process bridge. Every time you pan, zoom, or update in Vector/Hybrid mode, the architecture forces a resource-intensive re-calculation of every face, line, and edge, dragging down performance even on modern hardware.
IngeTrazo (The Modern Open-Source Challenger)
-
Underlying Architecture: Developed using modern Python and Qt / PySide6 frameworks, combining a high-performance geometric engine with a clean script-driven action API.
-
Code-Based Pros: Highly modular, lightweight, and agile. Its integrated design choice—building the sheet composer directly into the main application core—eliminates the laggy inter-app communication loop that plagues SketchUp and LayOut. This allows it to render scaled vector sheets smoothly.
-
Code-Based Cons: As a younger project built on a Python/Qt wrapper stack, it lacks the decades of deep, low-level performance tuning that legacy commercial C++ software has undergone for massive, hyper-complex enterprise models.
Future Outlook & Summary
While SketchUp remains safely entrenched due to its massive corporate ecosystem, its aging C++ core and rigid legacy structure (exemplified by the sluggishness of LayOut) leave it vulnerable to nimble, open-source alternatives. IngeTrazo’s modern, modular code foundation allows it to adapt rapidly to new tech—like AI automation and native cross-platform support—putting pressure on commercial giants to innovate faster or risk losing ground to agile software movements.
I’d like to add a few nuances…
SketchUp isn’t “old” just because its core is written in C++… and I believe it still outperforms IngeTrazo, especially since the latter still lacks optimization algorithms. It remains very modern today, and the fundamentals of 3D modeling haven’t changed much in decades.
What makes IngeTrazo interesting is that it is developed in Python—which is slower but more elegant and readable for an Open Source project—while deriving its performance from frameworks (written in C/C++) that handle 3D processing and high-quality rendering.
Python and Ruby are both developed in C/C++ and share key similarities, meaning that if you learn to program in one, it is easy to switch to the other. What makes C/C++ much faster is that it can be compiled into machine code; it operates closer to the hardware in certain respects, such as memory management.
I haven’t studied IngeTrazo’s source code in depth due to time constraints, but my impression is that we share the same mathematical foundations and basic algorithms (which are undoubtedly the same ones used in SketchUp), though some algorithms are not yet fully refined or are missing entirely. Still, it is a very interesting project that is already functional and continuing to evolve.
I’m on a Mac running safari, is it possible to do this with my setup?
Thanks for the technical comments on this one Alan4, its most appreciated.
GOOGLE-GEMINI also noted comparisons with the two programmes with regards to (Sheets), i.e. Layout. Please see below, your thoughts on this please…
SketchUp and IngeTrazo Sheet Layout
SketchUp’s companion application, LayOut, is notoriously sluggish, which directly relates to both its legacy programming architecture and its decoupled application design. LayOut operates as a separate binary that communicates back to SketchUp viewports. Every time you pan, zoom, or make a minor adjustment in “Vector” or “Hybrid” mode, LayOut forces a heavy, resource-intensive re-calculation of every single face, line, and edge across the boundary bridge. Because it relies on older, monolithic C++ foundational code that struggles with modern multi-threading, it frequently bogs down systems even on high-end hardware.
This performance bottleneck highlights a major advantage for modern challengers like IngeTrazo, which builds its sheet composer directly into the main application core using modern Python and Qt frameworks. By avoiding the laggy inter-app communication loop that plagues SketchUp and LayOut, IngeTrazo can render scaled vector views and updates much more smoothly.
Layout might be notoriously sluggish for some but for many it most certainly is not!
Bugs notwithstanding, Layout is a breeze for me.
we should ask gemini. 
2 Likes
First of all, it is not certain that IngeTrazo renders drawings using vector graphics. This point needs verifying; perhaps the framework handles it. That represents a significant development task (I have been working on exactly that for months…).
Next, the slowness issues in Layout have been resolved; there was a major display bug related to Retina screen support I guess. That leaves the vector display. It is inherently slower, but there seems to be an optimization issue somewhere. Exporting vector AutoCAD files from SketchUp appears faster than rendering vector graphics within Layout.
The Google-Gemini hypothesis is interesting. I don’t know if it’s true, as I don’t have access to Trimble’s internal development details. Personally, I would have duplicated the vector rendering algorithms in each application rather than sharing them… If that is the case (it seem absurd), it could explain the slowness. In any event, the issue isn’t simply due to there being two separate applications. That said, integrating everything into IngeTrazo is a very elegant and compelling approach.
The Windowizer extension is an interesting example of a script based on the APIs, but it isn’t truly parametric in the way Dynamic Components are. It resembles a Blender extension of the same name.
At this stage of development, that isn’t very important.
As it happens, I have developed a functional parametric engine. I could adapt a version of it if this project moves forward. I need to prioritize my own project. But it could be interesting to develop an extension that is compatible with DCs—and might even improve upon them!
And of course, I’m not the only one who has mastered the acyclic graph principle, who can develop an extension of this kind.
One more detail: I developed it in Python, but I don’t want to release it as open source. If I rewrite it in C/C++, not only could I keep the code proprietary, but I could also make it extremely fast and high-performance (unlike the current DCs, which are written in Ruby). Three compiled C/C++ libraries (for Mac, PC, and Linux) could be very easily integrated into a Python or even a Ruby script. So, it could become a universal function across multiple software applications.
This is AI hogwash, don’t believe everything you read. Gemini is regurgitating a mashup of terms it harvested from reddit backroom rants. There is no apples to apples comparison here, and much of the the below simply does not make any sense.
At this stage, it’s too early to weigh pros and cons. IngeTrazo’s main advantage is being open-source, giving it the potential to become the ‘Blender of Architecture’—provided the project is well-managed and gets the community backing Blender once did. However, launching an open-source project today comes with unique challenges. Years ago, companies like Autodesk and Maxon didn’t expect a small foundation to disrupt the industry standard for 3D artists. Today, companies like Trimble are far more watchful and will likely throw up hurdles to slow down competing tools.
I plan to reach out to the developer and coordinate community support to contact Ton Roosendaal. Getting his insights and backing could cut development time in half or more. It would also serve as a wake-up call to major industry players like Autodesk, Graphisoft, Trimble, Vectorworks, and Rhino—though Rhino is slightly safer given its perpetual licensing model.
I really hope that Sketchup plugin developers put an eye on this as well, my workflow relies heavily on extensions and i will be more than happy to buy again all the extensions I´ve bought throughout the years if they make them compatible with Ingetrazo.
Also contrary to what the guy who makes the videos of this program, as a native Spanish speaker, I don’t find it to be a good name, the creator used two words Ingeniero or ingeniería and Trazo which means stroke, we could think on more clever names.
I contacted them myself. They basically pointed me to the GitHub issues page for any bugs or suggestions, and they do respond there:
So at least from that side, they seem pretty open to feedback and community input.
At the rate things are going, you’d probably even be able to just build an extension for Blender that gives you the equivalent of a Sketchup overlay; using Blender as the engine, but presenting a Sketchup framework via extensions that makes the experience nearly identical.
Would probably take a lot of work though, but could probably help put more eyes on Blender by creating as close to an environment as possible and gradually open people up to the wider Blender environment workflows,
Go for it guys…
Speaking of alternatives, I’ve been following this project too. Different market obviously, but it’s interesting seeing how many alternatives to established software are starting to appear.
Hi endlessfix,
As previosuly noted, AI can sometimes paint a distorted picture.
I guess if I asked GOOGLE-GEMINI-AI where it attained its info from, would it provide an honest answer…
One other 
Yeah, if you think about it, instead of paying $500 a year for SketchUp, you could spend roughly the same amount on a year of Claude and ChatGPT at $20/month each and just start building something yourself.
With how capable these models are becoming, I’m pretty sure after a year you could come up with something pretty useful, especially if you only need it for your own workflow.