SketchUp Plugins · Technical Drawing · Automation · Freeware · 2026

If a plugin can do it, why should I?

There's an old saying about giving a difficult job to a smart but lazy person, because they'll probably find an easier way to do it. I don't remember who said it. But I think I understand the idea.

I enjoy designing things. What I don't particularly enjoy is doing the same repetitive work over and over again. Take technical drawings, for example. Imagine I've finished designing a table in SketchUp. The geometry is there. The dimensions are there. Everything is already modeled in 3D.

But then I still have to prepare the technical drawings. Top view, side views, perspective, dimensions, layout, and finally export everything into a PDF.

Wait a minute.

All that information is already inside the model. Why am I doing this manually? So I made a plugin.
Now I just select the model, click Generate Technical Drawing, and the plugin does the rest. Multiple views, complete with dimensions, exported into a PDF that's ready to send.

One click. Done.

Then my teammates saw it. "Hey, can I have that plugin?"
Apparently, I wasn't the only one who didn't enjoy doing things the hard way.

And that got me thinking. I've actually been making SketchUp plugins for two different reasons.
Some are designed to speed up the process. They help me work faster while I remain in control of the creative decisions.

Others follow a much simpler philosophy:

If a plugin can do it, why should I?
For tasks that are repetitive, predictable, and can be handled through a set of rules, I'd rather let the computer take care of them. Of course, sometimes I probably spend more time developing a plugin than it would take to finish the original task manually.

But that's only the first time.

The next time I need it, the plugin is already there. And the time after that. And the hundredth time after that.
Eventually, I decided to share some of these plugins for free on GitHub. Nothing particularly revolutionary. Just little tools I originally made to make my own work easier.

If they happen to save someone else a few hours, even better.
Maybe I'm lazy. But I prefer being lazy once.

The Eureka Moment

The information was already inside the 3D model. I didn't need to reproduce it manually. I just needed to build a tool that could turn it into a finished document.

Remote Rendering · Tailscale · MCP · Workflow Automation · 2026

Why should rendering take away my evenings?

I have a fairly powerful rendering PC at home. The problem is, I don't particularly enjoy spending my evenings sitting in front of it. Especially after working all day. Previously, I'd finish my 3D modeling at the office, then continue the rendering process at home. Setting up materials, cameras, lighting, and waiting for the final output.
It worked, but it was taking away my rest and family time.

Then I started experimenting with MCP, Tailscale, and headless rendering.

I connected my office workflow to my PC at home, scripted the rendering process, and prepared reusable templates for the settings I normally had to configure manually. Now I can send a rendering job from the office, let my PC at home handle the processing, and get high-resolution JPEG or PNG files back without having to sit in front of it.

But here's the interesting part.

While my computer is rendering at home, I can continue doing something else at the office.
Two jobs running at the same time, without me having to divide my attention between them.
I wasn't really multitasking. My computers were.

And perhaps more importantly, when I get home, I can actually spend my time at home.

One More Thought

My mind has a habit of jumping between ideas. Sometimes, in the middle of doing something, another idea suddenly appears. Before AI, I'd either write it down or simply forget about it. Which I always thought was a waste.

Now, when an idea comes along, I can ask AI to explore it, plan the technology stack, or start developing something while I continue working on whatever genuinely needs my attention.

Of course, I still need to review what it produces.
But at least the idea doesn't have to sit there waiting for me.
My mind still jumps around. Now my workflow can keep up with it.

The Eureka Moment

I didn't need to work on multiple things at once. I needed to let my computers work independently while I focused on what actually required me.

Data Security · AI Architecture · Local LLM · 2026

Does AI really need to know everything?

There's one thing that always comes up when introducing AI into a company: data security. And I think management has every reason to be concerned. I've seen how easily people upload entire documents or sensitive information into cloud AI services, simply because it's convenient.

But here's something I've been thinking about: does AI really need to know everything? Most of the time, it doesn't.

I use an approach I call distilled data. Instead of giving cloud AI the complete information, I give it just enough to perform the task. Sensitive details are separated, while identifiers preserve the relationships needed to reconnect everything later.

The results can then be stitched back together locally, using the original information that never needed to leave the controlled environment. For that part, I use Ollama with Qwen Instruct. Nothing particularly fancy. I don't need an extremely intelligent model just to reconnect information. I need something fast, accurate, and focused on a specific job, with deterministic checks where precision matters.

Of course, incomplete data isn't automatically safe. Information can still reveal things when combined, so the boundaries need to be designed carefully.

The idea is simple, though.
I don't need to give AI all my data to get useful work out of it.
Sometimes the trick isn't finding a smarter AI. It's figuring out how little the AI actually needs to know.

The Eureka Moment

AI doesn't need the complete picture to perform a specific task. Give it what it needs, keep sensitive context under control, and reconnect the results locally.

Creative Automation · MCP · Workflow Transformation · 2026

What if we stopped doing the work machines could handle?

I've always wondered why we still spend so much time doing repetitive work when we have perfectly capable computers sitting on our desks. The problem wasn't really the idea. I knew what I wanted to automate years ago. The technology just wasn't quite there yet. It's a bit like playing Age of Empires. You can have a great strategy, but you can't build certain things until you've unlocked the right technologies.

Then AI, MCP, and better application integrations came along. Suddenly, connecting Illustrator, SketchUp, and KeyShot into one workflow became possible. I started experimenting. What used to involve repetitive manual operations could now be handled through connected processes. One of my end-to-end experiments took less than an hour to set up. But the interesting part wasn't the automation itself.

It was realizing that we could rethink how the entire team works.

I don't want AI to replace my designers. I want my designers to stop wasting their abilities on tasks that don't really need them. And now that the technology is available, we can finally start building that kind of workflow. It's still evolving. But that's what makes it interesting.

The Eureka Moment

The technology was finally ready. The bigger opportunity wasn't doing the same work faster, but changing how the work gets done.