Blog I Co-Built a 3D-Printed Stand with Notion AI (And Never Opened Tinkercad)
I Co-Built a 3D-Printed Stand with Notion AI (And Never Opened Tinkercad)
How one photo and a chat with Notion AI turned into a print-ready 3D stand for my Maono AME2, from first sketch to documented project in a single conversation.
A few weeks ago I looked at my Maono AME2 sitting flat on my desk and thought: this should be angled towards me, lifted at the back, and my cables should not be fighting for space behind it. A custom 3D-printed stand would solve it. Normally that means an evening lost to Tinkercad, callipers, and regret.
Instead, I opened a chat with Notion AI (my custom agent, Sir Work-alot) and co-built the whole thing in one conversation. Here is exactly how it went, step by step, because the process surprised even me.
Step 1: One photo, one ask
I snapped a photo of the AME2 on my desk and sent it with a simple brief: design a table stand that angles the unit diagonally at height, give me an STL file, include the specs, and keep my cables out of the way.
The actual photo I sent. Messy desk and all.
No measurements from my side. The AI looked up the AME2's published dimensions (245 x 157 x 55 mm), checked where the ports sit, and came back with a full design: an 18-degree angled frame with a front lip, side rails, and an open rear window so every cable drops straight down.
Step 2: The first STL arrived in minutes
Not a sketch, not a description. An actual downloadable STL file, generated from scratch, with a rendered preview, full specs, and print settings.
Version 1: a one-piece angled frame. Looks great. One small problem.
Step 3: Reality check, instant redesign
The problem: the frame was 270 mm wide. My Bambu Lab P2S prints up to 256 mm. I told the AI, half laughing, that the thing was bigger than my printer.
One message later, the design was reworked into something honestly better: a minimalist two-cheek stand. No base plate, no full shelf. Just two small wedges with retaining rails, plus a 245 mm spacer bar that locks them at exactly the right distance. Largest printed part: 170 x 28 x 80 mm. Filament dropped from roughly 350 g to about 90 g, and the cable problem solved itself because the whole underside is open.
Version 2: two cheeks and a spacer bar. Prints on a single plate, no supports.
Step 4: Dumb question, patient answer
At one point I genuinely could not picture where the spacer bar went. So I asked, "where does the spacer go again, can you show me a picture?"
Seconds later, a rendered assembly diagram: both cheeks in position, the spacer bar highlighted in green, and a ghost outline of the AME2 sitting on top.
The assembly diagram. The green bar is the spacer. Instantly clear.
No judgement, no RTFM energy. Just the answer, drawn for me.
Step 5: It documented itself
Finally, I asked it to record the whole project in my Notion Projects database. It created a proper project page with the STL files embedded, both design versions, specs, print settings, an assembly checklist, and open checks (measure the unit before printing, test cable bend radius after).
That is the part that stuck with me. The conversation did not evaporate. It became a permanent, organised record in the same workspace where I run everything else.
What this actually felt like
Convenience is the wrong word, honestly. It is compression. An evening of CAD work, measuring, redesigning, and documenting collapsed into one chat while I stayed in flow. Every iteration was one sentence long: too big, show me, record this.
A few things I took away:
Start with a photo. Real context beats a perfect brief. The AI designed around my actual desk, my actual cables.
Iterate out loud. "This is too big for my printer" was all it took to get a genuinely better design, not a compromise.
Ask the dumb questions. A quick "show me" saved me from assembling it wrong.
Let it file the paperwork. Having the process logged in Notion means future me can reprint, tweak, or share this without digging through chat history.
I build Notion systems for a living, and I still underestimate how much friction disappears when the AI lives inside the same tool where the work lands.
What is next
The STL is sliced and ready. Next step is printing it on the P2S, and I will update this post with photos of the finished stand on the desk. If it survives a live session with my cable situation, it earns a permanent spot.
If you have a small annoying problem sitting on your desk right now, try describing it to your AI of choice before you open any design tool. You might skip Tinkercad entirely.
How I turned a broken Beeper extension into a local, cyberpunk command center that unifies all my chats — the decisions, the architecture, and the bugs that taught me the API.
What happened when I spent three hours on a PRD, then asked Claude to build the first Clicker Foundry prototype. Wins, tracing troubles, test prints, and lessons learned.
A recap of this week's Notion AI Office Hour — the four levels of AI usage, why Notion is your memory layer, how MCP really works, and when a worker beats an agent.