1 free creation, like a video, podcast or flashcards
Videos and podcasts up to 2 min
Upload 1 file, up to 8 MB
Try chat and quizzes on it
An AI 3D model generator that explains.Labeled, moving, and yours to turn around.
Ask for what you can't picture, like a heart valve or a four-stroke engine, or upload the chapter. Scholarly builds it, and you change it by asking.
In Beta. Runs in your browser, nothing to install. No credit card required.

What does this AI 3D model generator make?
Interactive 3D models that explain something: how the parts fit, what sits inside, what moves and why. Describe the subject in a sentence, or give it your PDF, notes or slides. Scholarly writes the model as a small three.js program with real geometry, lighting, labels and animation, then renders it from four views and fixes what looks wrong before you see it. 3D models are in Beta.
It is a model for understanding, not a game asset or a printable mesh. Best for anything you grasp faster once you can turn it around: a heart, a cell, a pump, a molecule, a network.
- What goes in: One sentence is enough. Or PDFs, Word files, slides, photos of notes, web pages, YouTube links and recordings. Free: one source up to 8 MB, 32 PDF pages. Premium: up to 10 sources, 300 MB, 1,000 pages. Laureate: 20 sources, 1,000 MB, 3,000 pages.
- What comes out: A labeled model in your browser that you orbit, zoom and set turning, with animation and up to six sliders and switches when they help. Share it as a live link.
- Changing it: Ask chat. Rewording a label or changing a color, a size or a speed is a small edit that uses no credit, and Undo steps back. A rebuild comes back as a draft you apply or discard.
- Credits and exports: A build uses 1 AI Creation Credit, and Free includes one. Premium gives 15 credits a week and Laureate 80, with every export: a 4K PNG, a transparent PNG, a turntable video and a .glb.
Cut the Earth open. Then drag it around. It runs in the page.
Move the cutaway slider to open the wedge, orbit, zoom or jump to a view. We wrote this small model by hand so it loads fast. It runs in the same viewer as the models Scholarly builds, which go into far more detail.
One sentence each. Real builds, in the chat that asked.
For nursing, medical and A&P students. Cut a structure open and see what sits behind what.

Try asking: A human heart cut in half to show the chambers and valves, every part labeled.
3D anatomy modelsFrom a sentence to a model in three steps. You say what to build. The building is ours.
1Describe it, or add a source
Sign up and the 3D model creator opens. Upload a PDF, slides or notes and pick the pages, or choose Start empty and describe what to build. Later, ask for a model in any chat.
2Say yes to the build
Scholarly suggests a model from your source, or follows your description. Nothing starts until you confirm the 1 AI Creation Credit. A build takes a few minutes, and you can leave the page.
3Turn it, change it, send it
Drag to look around, or jump to front, side or top. Ask chat for changes. Share a live link, or on a paid plan export an image, a turntable video or a .glb.
Bring your own material. The labels follow it.
Build from your chapter, manual or lecture notes and each part carries the name your source gives it. The model teaches what your course or your team actually says, and the words on screen are the ones you will be asked about.
- It reads the pages you pick, so the model stays on your subject.
- A photo of a textbook figure or a whiteboard works as a source too.
- Your sources stay private when you share the model.
- A video lecture, flashcards or cited answers can come from the same source.

Three kinds of 3D model, three jobs. Each is good at something.
Mesh generators such as Meshy
Turn a prompt or an image into a textured 3D mesh in under a minute, with STL, OBJ, FBX, GLB and USDZ export. Built for games, VR and AR, and 3D printing.
Ready-made 3D libraries
Polished models made by experts, ready to open. They show the standard version of a subject, labeled the library's way.
Scholarly
A model that explains, built from your sentence or your material and labeled in its words, with motion and sliders. Change it by asking, then make a lecture or flashcards from the same source.
What people ask before they build one.
How do I generate a 3D model with AI?
Sign up for Scholarly and the 3D model creator opens. Describe what you want to see, such as "a four-stroke engine, animated step by step", or upload a PDF, slides or notes and pick the pages. Confirm the build, which uses 1 AI Creation Credit, and the model appears in a few minutes. You can also ask for a 3D model in any chat.
Is the AI 3D model generator free?
Signing up is free, with no credit card. The Free plan includes one AI Creation Credit, and a build uses one, so you can build a model, view it and share it by link. Premium gives 15 credits a week and Laureate 80. Small edits use no credit, and if a build does not finish, the credit comes back. Downloads need a paid plan.
How is this different from Meshy and other text-to-3D tools?
Mesh generators such as Meshy turn a prompt or an image into a textured 3D mesh for games, AR and 3D printing. Scholarly makes models that explain: labeled parts, cutaways, motion and sliders, built from your description or your own material, and changed by asking. It exports a 4K image, a transparent image, a turntable video and a .glb file, but not STL or OBJ, so its models are not meant for 3D printing.
Can I make a 3D model from a PDF or my notes?
Yes. Upload a PDF, Word file, slides or notes, choose the pages, and Scholarly builds a model of what they describe, labeled with the names your source uses. Premium combines up to 10 sources in one model and Laureate up to 20. The Free plan builds from one source.
Can I turn a photo or an image into a 3D model?
You can use a photo as a source: a textbook figure, a whiteboard sketch or a page of notes. Scholarly reads what the image shows and builds an explanatory model of it. It does not scan an object from photos into an exact 3D copy.
Can I put the 3D model in PowerPoint, Keynote or a video?
Yes, on a paid plan. A 4K PNG, a transparent PNG or the 8-second 1080p turntable video goes into any deck or video editor. The .glb file opens in PowerPoint versions that support 3D models, and in Blender and web 3D viewers, with its shape and labels but without the animation. For Keynote, use the image or the video.
Can I edit the 3D model?
By asking chat. Rewording a label or changing a color, a size, an animation's speed or the background is a small edit that uses no credit, and you accept it on a card. A bigger change is a new build, which comes back as a draft you preview, then apply or discard. Undo brings back the previous version. There is no sculpting or dragging parts by hand.
How accurate are the models?
They are teaching models, not measured to scale or engineering-grade. Before you see a build, Scholarly renders it from four views and fixes errors and labels that cover each other. Built from a source, the labels use its names. The AI can still get something wrong, so check the labels before you study from a model or show it to a class.
How long does a build take?
A few minutes. Keep working or leave the page while it builds, and the model is there when you come back.
Can I share a 3D model, or open it on my phone?
Yes. Send a link and the other person gets the live model to turn around, not a picture of it. Your sources stay private: their pages do not appear for people you share with. On a phone, a model opens full screen with View in 3D.
Start from what you have. Or turn the same source into a lecture.
A build is one credit. A small edit is none.
Pricing shown in USD by default. Local currency may be available in app or at checkout.






