Open navigation menuClose navigation menu

Make a 3D model from your PDF

Open a manual, protocol, spec, or study guide in 3D. Scholarly's AI reads the pages and builds a labelled 3D model of what they explain, then you refine it in chat and export it for slides and video.

Sign up for freeHow 3D models work

Beta. A full AI build uses 1 AI Creation Credit, and you confirm it first. Small chat edits are free.

A labelled 3D model built from a PDF in Scholarly, with the chat beside it

Quick answer

Can AI turn a PDF into a 3D model?

Updated October 2026 · 3D models are in Beta

Yes, for explanation rather than manufacturing. In Scholarly you open a PDF as a 3D model; its first pages appear in 3D. Ask the chat beside it for a model of what the document describes, and Scholarly's AI writes the model as real three.js code, with geometry, materials, lighting, labels from the source, and animation where it helps. You preview the draft and apply it or keep your current model. It does not convert drawings into CAD files, STL, or printable meshes.

Most "PDF to 3D" tools are converters: they extrude a floor plan or trace a technical drawing into a mesh. That is a different job. Scholarly is for documents that explain something: the anatomy behind a nursing procedure, the parts of an infusion pump, the tiers of a cloud architecture, or the stages of a gear train, where a model you can turn around lands faster than another page of text.

Because the model is built from your document, the labels say what your document says. The AI reads the pages before it builds, so a label like "Mitochondrion" or "Load balancer" comes from your source, not a generic template. And because the model is real three.js code rather than a few stock shapes, it can be a plant cell cut away to show its organelles with leader lines and readable labels, or a gear train that actually turns.

The result is a working draft. Preview it and apply it, or keep the model you have. Small edits in chat (a label's wording, a colour, a size, an animation's speed, the background) are free and appear on an approval card, and Undo brings back the previous version. On a paid plan, export a 4K PNG, a transparent PNG, a 1080p turntable video, or a .glb file that opens in PowerPoint and Keynote.

How it works

From PDF to 3D model in four steps

One PDF in, one labelled 3D model out, and you approve the draft before it replaces anything.

  1. 01

    Open a PDF as a 3D model

    Pick a PDF from your library or upload one. Its first pages appear in 3D, ready for the model to be built from them.

  2. 02

    Ask for a model of the subject

    In the chat beside the model, say what to build, for example "a cutaway of the system in section 2, labelled". Confirm the build, which uses 1 AI Creation Credit.

  3. 03

    The AI builds it from your pages

    Over a few minutes it reads the pages and writes the model in three.js, with geometry, materials, lighting, labels, and animation.

  4. 04

    Preview, apply, refine, export

    Preview the draft and apply it or keep your current model. Make free small edits in chat, then export an image, a turntable video, or a .glb on a paid plan.

Use cases

What people turn PDFs into

Documents that describe structure, equipment, anatomy, or how parts relate make the best models.

Equipment and technical manuals

Show how the components of a machine, a pump, or an HVAC unit fit together, each part labelled from the manual.

Clinical and nursing material

Turn a protocol or anatomy chapter into a labelled cutaway you can rotate while you explain it.

Certification study guides

Model the cloud architecture, network tiers, or anatomy your exam outline keeps referring to.

Engineering and system specs

Build an explanatory model of a mechanism or a system from its spec, with moving parts animated.

Training decks and videos

Export a 4K or transparent PNG for a slide, a turntable MP4 for a video, or a .glb for PowerPoint and Keynote.

Research papers

Turn an experimental setup or a described structure into a model you can orbit and show.

Try asking

Requests that work well on a PDF

Point the AI at pages or sections and say what the model should show.

  • Build a 3D model of the system described on pages 3 to 6, with every part labelled.

  • Make a cutaway of the cell in this chapter, with leader lines to each organelle.

  • Model the cloud architecture in this guide, one tier above the next.

  • Animate the gear train from section 2 so it turns.

  • Rename the label on the pump housing to match the manual.

  • Make the background white so I can export it for a slide.

Beta

What it can and can't do today

3D models are in Beta. Here is exactly what a PDF build does and doesn't do.

Works today

  • Builds from one PDF; its first pages appear in 3D and the model is grounded in what they say.
  • Real three.js geometry, physically based materials, lighting, and animation, not a handful of stock shapes.
  • Orbit and zoom, perspective, front, side, and top views, frame all, and a turntable.
  • On paid plans, export a 4K PNG, a transparent PNG, an 8-second 1080p turntable MP4, or a .glb file.

Not supported

  • Not CAD-accurate or engineering-grade. Models explain how something works; they are not measured designs.
  • No STL or OBJ export, and no 3D printing. Exports are images, a turntable video, and a .glb file.
  • No manual 3D editing tools: no sculpting and no dragging parts. You change a model by asking chat.
  • A model starts from one PDF or from an empty chat. For notes or slides, save them as a PDF, or describe the content in chat.
  • It's a Beta: preview every draft before you apply it, and check the labels before a model goes in front of a team.

FAQ

Frequently asked questions

How do I turn a PDF into a 3D model?

Create a 3D model from a PDF in your Scholarly library or a new upload, and its first pages appear in 3D. Then ask the chat beside the model to build a model of the subject and confirm the build. Over a few minutes, Scholarly's AI reads the pages and writes the model as three.js code. Preview the draft, then apply it or keep your current model.

Can I export the 3D model to PowerPoint, Keynote, or as STL?

On a paid plan you can export a 4K PNG, a PNG with a transparent background, an 8-second 1080p turntable video (MP4), and a .glb 3D file that opens in PowerPoint, Keynote, Blender, and web 3D viewers. There is no STL or OBJ export and no 3D printing, so for a printable or manufacturing mesh, use a dedicated CAD tool.

Does it convert 2D drawings or floor plans into 3D?

No. It doesn't trace or extrude drawings, and it isn't CAD. It reads what the document explains and builds an explanatory 3D model of that subject, with labels from the pages.

Which pages of the PDF does the model use?

The PDF's first pages appear in 3D, and the model is built from what they say. For a long document, save the chapter or section you care about as its own PDF so the build stays focused on it.

What does a build cost?

Signing up is free. A full build uses 1 AI Creation Credit; you confirm it before it starts, it runs for a few minutes, and then you preview the draft. Small chat edits, such as a label's wording, a colour, a size, an animation's speed, or the background, are free and appear on an approval card. Exports need a paid plan.

How accurate is the model?

Labels come from your PDF, but 3D models are in Beta and they are explanatory, not CAD-accurate or engineering-grade. Use the model to explain, preview the draft before applying it, and fix anything off by asking chat. Undo brings back the previous version.

Open your PDF in 3D

Sign up, open a PDF as a 3D model, and ask for a model of what it explains.

Sign up for freeHow 3D models work

Beta. A full AI build uses 1 AI Creation Credit; small chat edits are free. Exports need a paid plan.