This is a new thread pulled out of a digression elsewhere. So some of this initial post is a repeat of that.
ChatGPT can be part of a workflow to 3D models but won't (currently) generate the model. Note that ChatGPT is quite capable of being untruthful (and frequently is). If challenged it will happily own up to it and, while this is amusing, it can be frustrating.
It's handy for generating the images that other AIs can use to produce models though.
Meshy, Hitem3D and Tripo (and others) can produce models. Below is the ChatGPT image when asked for a 30s female passenger standing waiting for a train. You can refine this as needed by asking for whatever suits. Using this image in the other AIs produces the 3D models illustrated. These are all high poly watertight printable models. Not perfect by any means but perfectly acceptable in scales less that 1:64 and easily refinable in larger scales.
This was all done with free/demo versions and one could viably get to a printable figure in 5 minutes.
This is far from the last word on the subject. Realistically I have only dipped a toe in the water and I anticipate that in coming months this early exercise will look inept. Notwithstanding that I've gone from a process that took days to one that takes minutes and virtually no skill. It has issues, particularly as it is not amenable to tight control, but is still extremely useful.
I've continued with printing the Hitem3D version and will report back. I like the Meshy version more, but it is not available for free download. I'm not averse to subscribing to one of these services, but there are a lot of them and deciding which to patronise is a challenge.
Timely discussion Lawrence!
I have just been investigating the exact same thing for making printable figures/ items from historic photos. I have only played a bit with Meshy's free version, but was considering a subscription if it works well. I had not heard about Hitem3D yet, so thank you for the introduction. Looking forward to your results from your first prints.
Quote from: Rail and Tie on February 17, 2026, 04:34:29 PMTimely discussion Lawrence!
I have just been investigating the exact same thing for making printable figures/ items from historic photos. I have only played a bit with Meshy's free version, but was considering a subscription if it works well. I had not heard about Hitem3D yet, so thank you for the introduction. Looking forward to your results from your first prints.
Thanks - I assume others are looking. The principle is dead easy, but I'm interested in anything that improves the process, makes it more controllable or speeds it up.
Generic people are useful, and fairly easy I think. What makes custom figures more fun is the ability to create custom cameos, which I suspect will be trickier.
So here she is painted.
She doesn't stand cruel enlargement but at just over an inch high she's amazing. I'm fairly new to figure painting and while I think I have a grasp of the basic techniques my execution can be wayward. My first attempt was a bit rough, so this is a partial rework.
The facial detail is there. It's beyond me to paint as such, but some washes have highlighted the detail enough to give a good impression.
There's a few hours in this paint job and, while I think I could do better, I don't actually need this figure so I'm going to leave her as she is and move on to some stuff I do need.
That turned out very well. The figure detail is spot on, especially for the scale. I have been trying to find a way with the AI stuff to exaggerate the creases/folds in fabric which can help painting in small scales. No luck yet, but will be trying some of this this weekend if I get some time.
Quote from: Rail and Tie on February 18, 2026, 10:58:56 AMThat turned out very well. The figure detail is spot on, especially for the scale. I have been trying to find a way with the AI stuff to exaggerate the creases/folds in fabric which can help painting in small scales. No luck yet, but will be trying some of this this weekend if I get some time.
I was pleased!
This AI route seems to be quite good at such detail. However, painting in detail works well as well. The crease/seam on the back of her jacket is painted in, not modelled. This is the established idea of using a base tone, base+black and base+white. The idea is to paint in creases etc using the lighter/darker shades of the base. It works very well. It helps if the detail is there, but where it isn't you can easily add some.
I've moved on a bit. I purchased subscriptions to Meshy and ChatGPT and I'm developing a workflow to produce specific figures that I'll write up at some point. What I want to achieve (and I'm basically there) is to start with an image (generally a period pic in my case) and to reproduce that figure convincingly in my desired pose without a lot of work. It's pretty simple and quick, but the AI steps take time (a minute or two each iteration), and it can take a few goes to get something that I'm happy with. Even so, it's realistic to be ready to print within 30 minutes.
Quote from: Rail and Tie on February 18, 2026, 10:58:56 AMI have been trying to find a way with the AI stuff to exaggerate the creases/folds in fabric
It may be that one AI will do this better for you than others. Hitem3D seems to generate sharp features that catch paint more than something like Meshy which is generally smoother and produces nicer details (in my opinion). If you are working in HO then detail levels are probably moot as all the AIs will give you more than you can usefully print/paint.
Thank you for what is becoming a very useful clinic on AI for modeling. -- Russ
A very interesting thread!
I think a hybrid approch would be the way to go for me. I need figures for my workshop diorama. I have some period images that I can use as a starting point. This fella has the right outfit, but the pose is a bit passive:
mann.jpg
Then I used ChatGPT to get him working on the lathe:
arbeider_2.png
Clothing and details seems to have preserved pretty well. Regarding the lathe, I feel that ChatGPT is not yet up to the task...
I played a little with my Hitem3D free credits during my lunch, break and things indeed look promseing:
3D-Mann2.png3D-Mann.png
All in all this little excercise took under 30minutes, including writing this post. With a little tweaking both of the images and the 3D model, this could really be something.
But really bad news for companies for Modelu...
I ran out of files, but as you can see, the Overalls of the worker has turned into a jacket during the prosess. Here is a screen dump of the .stl:
mann_stl.png
So I went back to ChatGPT to make a better starting image. I will look more into this workflow, but It seems like the best route is to ma a set of images in ChatGPT that shows the figure from more sides. For now, I had only time to make the overalls more detailed:
forbedret_kjeledress.png
What a weird time to be a modeller...
ellevill.png
Hi Hauk, you've latched on to my latest approach.
I've subscribed to ChatGPT and Meshy, which does open up possibilities and speed things up quite a lot.
My broad approach now for a specific figure is as follows (and close to yours) - but it does vary a bit case by case. For generic figures it can be better to give the software more free rein as you get a smoother result with less artifacts.
- Find a reference image
- Use ChatGPT to isolate the figure, repose and generate multiple views. Multiple views may not always be needed, but where the unseen side is not obvious from the original image, additional images may provide the data.
- Use the references in Meshy. Try a single view first, or the multi (paid option) view if needed
- Meshy gives you 4 free retries. This is useful as they differ a bit and one is generally superior to the others for my purposes
- Export stl without resize (important to preserve poly count)
- Tweak in Blender if required. Typically adding a bit of additional depth to feet is useful and some simple sculpting to touch up details if needed.
Hi Hauk
Based on your last reference image, Meshy delivered the attached. It took longer to sort the images for this post and write it than it did to produce the models! Each model is 17-18MB, and nearly 400,000 polygons, so plenty of data. At 1:64 they would be great. I think they would be good in larger scales too. More detailed than many offerings, but some of the artefacts might be more obvious.
The first view is in my slicing software, which doesn't always render smoothly. The other two images are the same figures in another viewer.
As you can see each try differs a bit. The software seems to struggle with headgear and the caps could do with some some additional sculpting.
Experience suggests that these would print well. They generally look better too as these viewers tend to distort the perspective a fair bit.
I'm happy to share the models if they are useful to you.
Awesome work Hauk and Lawrence! You have set me up nicely for success. It seems like this is taking off in the MRR world as well. I see that Doug over at Foscale models has released his first figures. I suspect using this exact same method in his "Humanity to Scale" series.
https://fosscalemodels.com/collections/humanity-to-scale
Also Bernard over at miniprints offers a 3D figure from a photo which I am sure uses meshy.
Things are moving pretty fast in this AI space.
I'd better learn to deal with AI because, in a couple of years, it may be possible to design a complete locomotive (kit) and all kinds of other goodies. I suspect, though, the whole process (from photos and/or drawings to AI to 3-D printing to cleanup and assembly) may result in some of us doing computer work and others focusing on assembly and finish. However things progress, it's inspirational. -- Russ
Quote from: finescalerr on February 20, 2026, 01:28:23 PMI suspect, though, the whole process (from photos and/or drawings to AI to 3-D printing to cleanup and assembly) may result in some of us doing computer work and others focusing on assembly and finish. However things progress, it's inspirational. -- Russ
That ship has sailed. The great thing about AI (for this application at least) is that there is MUCH LESS computer time required.
For things like figures where dimensional accuracy is not tremendously important it's really useful.
FWIW I recently ordered several figures from Modelu.
I was very impresario the quality, service and ability to print in 1/16th (who models in That scale?)
Sorry if this has nothing to do with the thread.
I ordered new cutters to remove the supports.
MJinTN
The likes of Modelu (mostly) use 3D scanning to generate models. Good poses and clothing drape. If you want to do this you need a set up as well as having to find models and clothing to suit. It's quite feasible but, especially if you want period, it's awkward. I've never been all that impressed with the resolution of their faces, but I haven't studied recent offerings with any rigour.
The way I've done it the last few years is via digital sculpting. Superb detail and resolution, but poses and clothing drape require skill and I'm not that good at it. It takes time too - a lot.
The AI based results are as good as Modelu I think in terms of pose and clothing. Face detail probably on a par depending what AI you use. Not as sharp as sculpting, but better than I can paint in a small scale. But it's cheap and super quick and you can work from old images. Most importantly you get a unique figure that is just what you need in less time than it takes to navigate Modelu's website. As I'm about as far from the UK as it's possible to be I don't have to wait on shipping either.
Even if the AI result does not fully meet your needs, it would be a good start for some additional sculpting and you could always graft a high resolution face onto a figure. I've felt no need to do that, but in 1/16th it would be an option.
Like everything, it is not a technique that will suit everyone. But things that are cheap, easy, fast and excellent quality tend to find favour.
Quote from: Lawrence@NZFinescale on February 20, 2026, 09:55:55 AMHi Hauk
Based on your last reference image, Meshy delivered the attached. It took longer to sort the images for this post and write it than it did to produce the models! Each model is 17-18MB, and nearly 400,000 polygons, so plenty of data. At 1:64 they would be great. I think they would be good in larger scales too. More detailed than many offerings, but some of the artefacts might be more obvious.
The first view is in my slicing software, which doesn't always render smoothly. The other two images are the same figures in another viewer.
As you can see each try differs a bit. The software seems to struggle with headgear and the caps could do with some some additional sculpting.
Experience suggests that these would print well. They generally look better too as these viewers tend to distort the perspective a fair bit.
I'm happy to share the models if they are useful to you.
It looks like Meshy gives slightly better results than Hitem3D, the facial expressions look better, and the hands also looks healthier on your version.
I would be very happy to make some test prints from the models!
It is really exciting that such good results can be achieved so easily and relatively cheaply.
As excited as I am about these figures, I still think the technology has quite a way to go when it comes to things like engines and other rolling stock. Or a lathe for that matter.
ChatGPT and other Ai software has an insane amount of photos of humans in their training data. And the variations are relatively minor. And let's remember that even as good as the figures we are looking at are, they are still generic.
But I might underestimate the technology. I guess it is just a matter of time before Lawrence does a test with a specific technical object like a piece of machinery.
Thinking a little deeper, I realize that I am deifting towards describing 3D-scanning.
So what's really new here is what Lawrence already have pointed out: The simplicity for the user, and the possibility to model subjects you don't have access to outside of historical images and descriptions. And things that exist only in your imagination. (like a lathe operator going bananas!)
I think the technology we are discussing best can be be summed up in the Science Fiction Writer Arthur C. Clarkes famous quote: Any sufficiently advanced technology is indistinguishable from magic.
Quote from: Hauk on Today at 03:07:23 AMBut I might underestimate the technology. I guess it is just a matter of time before Lawrence does a test with a specific technical object like a piece of machinery.
Thinking a little deeper, I realize that I am drifting towards describing 3D-scanning.
So what's really new here is what Lawrence already have pointed out: The simplicity for the user, and the possibility to model subjects you don't have access to outside of historical images and descriptions. And things that exist only in your imagination. (like a lathe operator going bananas!)
I think the technology we are discussing best can be be summed up in the Science Fiction Writer Arthur C. Clarkes famous quote: Any sufficiently advanced technology is indistinguishable from magic.
We stop improving when we think we are good enough. I've not been tempted to use AI on 'hard' subjects as I'm quite happy to use CAD for those. Some of the AI marketing suggests it would be OK, but as modellers we are after accuracy whereas the CGI crowd wants something that looks believable. The other problem is that for figures it is relatively easy (for a human or machine) to 'create' the unseen views. For something like the lathe you actually need some specific knowledge to do that. So I can see AI creating the lathe with a body of data, but not from a single image - unless the accessible database becomes large enough.
The other consideration is how you define your hobby. I like CAD modelling so I'm not looking for change. I'm a paint by numbers artist, so my sculpting is never likely to be great so AI is a godsend for figures.
A friend of mine's father was stationmaster at Rewanui when he was a child. I have photos of him around the garden, so one of the things I wanted for the layout was to capture him at that time with the family dog as a cameo. AI has allowed me to reproduce both child and dog very quickly - something I doubt I ever could, or would, have achieved otherwise. The result below is superb, other than the facial features which are good, but not accurate. The source image is average, and the model tiny in 1:64 so this is not a problem for me as the model is considerably better than I can reproduce or paint. I think Hitem3D offer a portrait mode that might produce a better facial result if that was needed.
The creative possibilities are infinite. Modellers will be able to recreate historic events and personal memories on an unprecedented scale.
A small number of extremely skilled miniature sculptors have always been able to achieve this, but us mere mortals had to settle for replicas of their work, and we had to accept that the small folks populating our scenes showed up on an awful lot of dioramas and layouts.
As a prototype modeller I strive for authenticity. Originality is also important for me, I want to model prototypes that hasn't been modeled to death.
Conventional CAD combined with advanced techniques like photoetching, CNC-machining and 3D-printing has made it possible for me to scratchbuilt almost anything man-made related to my prototype of choice.
Like Lawrence, I'm more of an engineer than an artist, so sculpting figures from scratch was never an option. Figures based on 3D-scannned people was something I considered trying. But as we see, technology has moved to the next level before I could catch up!
As exiting all of this is, I think it is well worth noting that all the recent revolutions is in the software. When it comes to hardware, the developments are more incremental. No doubt 3D printing was revolutionary, but we still have no color 3D printers that can rival handpainting and weathering. There are no high resolution metal prints available for hobbyists.
I guess my point is that there is still a place for craftsmanship and creativity in the hobby.
But this is not your father's hobby...