The argument, briefly
Every other media type has a portable document.
Text has PDF. Images have JPEG. Video has MP4. Audio has MP3. Each of those formats won not because it was the most technically elegant option available, but because it solved a specific human problem: send this to someone, anywhere, and trust that they can open it.
3D has no such format. We have plenty of 3D models — OBJ, STL, FBX, STEP, GLB, USDZ. Each is excellent at what it does. None of them are documents.
A 3D model is geometry. A 3D document is geometry plus everything you need to share, view, and understand it on its own. Metadata. Presentation settings. AR behavior. A viewer that runs without installed software. An archival shape that will still open in ten years.
This is the case for why.
What the document world figured out
The PDF story is worth telling briefly, because the lesson is the whole point.
Before PDF, sharing a formatted document was a coin flip. The recipient needed the same software you used. The same fonts. Often the same operating system. A Word document from one machine to another could render with the margins blown out, the typography substituted, the headers in a different place. The document existed, but the experience of the document didn't travel with it.
PDF solved that by reframing the problem. A PDF wasn't a working file. It was a finished artifact — the document as it was meant to be seen, packaged with everything needed to render it the same way everywhere. Fonts embedded. Layout fixed. Viewer freely available. One file, opens anywhere, looks the same.
The lesson is that the format won by being self-contained, presentation-stable, and viewer-agnostic. Not by being the most powerful editing tool. Not by being the deepest technical specification. By solving the one problem that mattered: I need to send this and have it work.
That problem is unsolved in 3D today.
The 3D PDF attempt — and what we can learn from it
The 3D world has tried this before. Adobe launched Acrobat 3D in 2006, with 3D content embedded directly in PDF files via U3D and later PRC. For a moment in the mid-2000s, it looked like the document format would simply absorb 3D the way it had absorbed forms and embedded images.
It didn't work. By around 2010, Adobe had handed off the 3D capability to third parties — Tetra4D and Tech Soft 3D — and the feature has lived in a kind of permanent twilight ever since. Modern browsers' built-in PDF viewers don't render 3D PDF content at all. Mobile devices ignore it. The dependency on installed Adobe Reader (or its successors) became the trap that killed the format: as the world moved to browsers and mobile, 3D PDF stayed tied to a desktop application most people had stopped opening.
The technical concept wasn't wrong. The execution was right for 2006 and wrong for everything after. A portable 3D document that depends on a specific installed viewer is portable in name only.
The lesson isn't don't try. The lesson is the viewer has to come with the file, and it has to run where people already are — in browsers, on phones, without installation.
That wasn't possible in 2010. It is now.
What's changed since 2010
Four things changed between 2010 and today, and together they make a real portable 3D document possible for the first time.
WebGL matured into a real rendering platform. By the mid-2010s, every major browser shipped hardware-accelerated 3D rendering. The question stopped being "can browsers render 3D?" and became "what do we render?"
glTF became the JPEG of 3D. The Khronos Group released glTF 2.0 in 2017, adding physically based rendering and the binary GLB container. The format was designed from the start to be runtime-friendly — small files, fast load, vendor-neutral. It became an ISO/IEC International Standard in 2022. For the first time, 3D had a true delivery format with the same character JPEG has for images: ubiquitous, efficient, broadly supported, royalty-free.
Mobile AR became a browser feature. Apple launched AR Quick Look at WWDC in June 2018 (released publicly with iOS 12 in September 2018), letting any USDZ file linked from a web page launch directly into AR on iPhone and iPad. Google followed with ARCore-based scene-viewer support on Android. For the first time, augmented reality stopped being a native app's feature and became a browser's feature.
Progressive Web Apps made offline viewers real. A modern browser can install a web application that runs without network access, caches its own assets, and behaves like a native application — but with no app store, no installation friction, and no platform-specific build pipeline. A 3D viewer can now be a web page that works offline.
None of these four things existed at scale in 2010. By the early 2020s, all four were production-ready. The pieces of a portable 3D document were sitting on the floor waiting for someone to assemble them.
Why GLB alone isn't the answer
GLB is excellent. It's the right model format. It belongs inside any portable 3D document — including p3DF™.
What GLB doesn't do is carry the document layer. It doesn't tell you what the model is, who made it, what project it belongs to, what SKU it represents, how the viewer should light it, what AR placement should look like, or which viewer to use. It's geometry and materials. By design, narrowly scoped, and that's exactly why it has been adopted so successfully.
But a GLB alone is to a 3D document what a TIFF is to a photo gallery, or a TTF is to a publication. The asset is there. The context isn't.
Trying to handle that context with a folder of related files — a GLB, a JSON, some images, maybe a viewer.html — doesn't survive the trip. Files get separated. Viewers get versions that don't match. Three years from now, the folder is incomplete and nobody remembers which viewer was the right one.
The document moment is the moment you bundle all of that into one thing that travels together.
What a portable 3D document actually needs to be
Five requirements. Each one is non-negotiable.
Self-contained. Everything needed to view, interact with, and understand the model is inside the file. No external dependencies. No "open this in Application X." No "now download the companion viewer." If the file works, it works alone.
Viewer-included. The viewer is part of the document. Not linked, not referenced, not assumed to exist on the recipient's machine. Included. The format must carry the means of its own rendering.
Browser-native. The viewer runs in any modern web browser, on any modern device, without installation. The browser is the universal runtime; anything that requires more than the browser has already lost.
AR-capable. Mobile AR isn't a bonus feature. It's the reason this technology is interesting now in a way it wasn't fifteen years ago. A portable 3D document must launch directly into AR on the platforms that support it — iOS Quick Look, Android scene-viewer — with no app, no install, no friction.
Archival. The file must still open in ten years. The format must not depend on any single company's product roadmap, any single platform's API, or any single server staying online. PDF won partly because a PDF from 1996 still opens. p3DF™ must aim for the same property.
p3DF™ is built to meet all five. The format bundles a GLB and USDZ together with manifest metadata and an embedded HTML5 viewer in a single ZIP container — the same container pattern used by EPUB, DOCX, USDZ, and 3MF, all of which have proven that approach durable across decades.
What changes when the format exists
The reason any of this matters isn't the file format itself. It's what becomes possible when the format exists.
Sales teams send 3D the way they send PDFs. A cabinet maker emails a client a p3DF™ of the proposed island. The client opens the email on their phone, taps the file, walks around the island in AR in their actual kitchen, and replies with sign-off — all without an app, an account, or a phone call. The friction that has kept 3D out of routine business communication disappears.
Customer deliverables include 3D as standard. A contractor finishes a project and hands the client a folder of documents at closeout. Today that folder contains PDFs of plans, warranties, and product specs. Tomorrow it contains those plus a p3DF™ of the as-built model. The client keeps it. Years later, a future contractor opens the same file to see exactly what's behind the wall.
Archives include 3D as a first-class format. Museums, libraries, government records, project closeout files — all the institutions that have been struggling with what to do about 3D get an answer that matches what they already do with documents. A p3DF™ is a document. Treat it like one.
Manufacturers ship product data as 3D documents. Spec sheets become p3DFs. Catalog entries become p3DFs. The customer can rotate, inspect, place in space, and read the metadata — all from one file.
3D becomes referenceable. Document workflows assume documents can be filed, searched, attached to records, embedded in other documents. 3D has been outside that workflow for thirty years because 3D wasn't a document. When 3D becomes a document, it joins the workflow.
None of this requires anyone to adopt new habits. It requires the document to exist in a form that fits the habits people already have.
What's next
p3DF™ exists today. It is shipping in ModelLink 3D, the reference implementation. Every model uploaded to ModelLink 3D can be exported as a p3DF™ and shared as a portable document. The technology works. The format is real. The trademark applications are filed. The path is open.
What happens from here depends on how widely the case lands. Formats win through use. PDF didn't win because Adobe decreed it; PDF won because enough people sent enough PDFs that the format became the default expectation. The same path is open for p3DF™.
If you build 3D content for a living, you already feel the gap this format fills. If you send 3D to clients, you already know what the workflow needs. If you archive 3D, you already understand why the document layer matters.
The conditions are right. The format is ready. The argument is the one PDF made forty years ago, and the answer to whether it works in 3D is the same now as it was for text then: it works when enough people decide it should.