Getting a room onto a headset takes one of three routes. Capture in a platform that ships its own VR mode, view a capture natively on Apple Vision Pro, or export a model and load it into a VR platform yourself. Which fits depends on the headset your audience owns, so settle that before scanning.
The Three Routes
What People Mean by VR Walkthrough
Is a 3D tour the same as a VR walkthrough?
The phrase covers three different things, and choosing tools before separating them is how projects go wrong.
A 3D tour runs in a browser on a laptop or a phone. You click through a space, and it feels like a slideshow with spatial connections between the frames. Most estate agency listings are this.
A virtual walkthrough usually means moving through a captured model on a screen, sometimes following the path the person walked while scanning. Our Virtual Walkthrough feature works this way: it retraces your capture path with high-resolution photos placed where you took them, and it exists to verify conditions remotely rather than for immersive presentation.
A VR walkthrough puts the space on a headset. Two displays, one per eye, tracking your head as you turn. That is a different technical requirement, and it is the only one of the three where the hardware your audience owns decides what you can build.
If your client will view this on a laptop, you don't need VR or the export work that comes with it. If they own a headset, keep reading.
Route One: Capture in a Platform That Ships VR
Which tool gives a VR walkthrough with no extra work?
Matterport is the shortest path here if your audience is on Quest, because the VR mode ships with the platform rather than being something you assemble.
Their VR Mode runs on WebXR, which means no app to install. You open a Space in the headset’s browser and select Enter VR. Movement is free rather than restricted to preset waypoints, 360 Views behave as they do on a screen, and the mode is available across all Spaces. Meta’s Web Launch tooling adds a View in VR button that sends a URL from a phone straight to the Quest browser, removing the step where someone has to type an address in a headset.
A couple of practical notes come with it. The Space needs to be set to public before it can be viewed in VR. And although the mode runs on WebXR, Matterport says it has not tested headsets beyond the Oculus line, so treat Quest as the supported target.
If a Quest-based VR walkthrough is the deliverable and nothing else about the project pulls you elsewhere, this is the answer.
Route Two: View a Capture on Apple Vision Pro
On Apple Vision Pro, you can, without any conversion step. Polycam runs natively there, requiring visionOS 1.1 or later, and captures made on an iPhone or iPad open in the headset from the same library.
You explore a captured space in the headset rather than presenting a designed VR experience with authored navigation, and there is no equivalent native app for Quest today. For a designer or contractor who wants to stand inside a scan of a room, it does the job with no export, no conversion, and no third-party platform. For a client-facing tour on whatever headset the client happens to own, it does not.
The practical constraint is the ecosystem. This route works if you and your audience are on Vision Pro, and does nothing for anyone on a Quest.
Route Three: Export and Bring Your Own Platform
Start by exporting a mesh. GLB is the usual target because it packs geometry and textures into a single file and most VR platforms accept it. On our plans, GLTF exports on the free tier, with OBJ, FBX, DAE, STL, and USDZ arriving on Basic and above. GLB and GLTF are the same format in binary and text form, which is why an export labeled GLTF often arrives as a .glb file.
Then choose a host platform. Social VR spaces and collaboration spaces accept uploaded models, and game engines give you full control at the cost of building the experience yourself. The platform, not the scan, determines what the walkthrough feels like.
Expect to lighten the model first. VR renders a separate view for each eye at a high refresh rate, and a raw capture mesh is usually heavier than a headset can drive smoothly. Decimating the mesh, cropping to the area you actually need, and checking the scale before you export all save time later. A capture that runs fine on a laptop can stutter badly in a headset, and the fix is almost always fewer polygons, not a better headset.
Preparing a Capture That Works in VR
Capture the whole envelope, ceilings included. A viewer in VR can look anywhere, and ceilings are the surface people skip most often when scanning, so a hole above your head is far more noticeable in a headset than on a screen.
Get the floor properly too. Anywhere the viewer stands, they may look down, and gaps in floor geometry break the illusion faster than wall artifacts do.
Watch reflective surfaces closely. Mirrors and glazing produce phantom geometry, which is merely odd at desk scale and genuinely disorienting at full height.
Crop hard before you export, since neighboring buildings, sky, and outdoor noise all cost frames and none of them help.
Check the scale before it reaches a headset. A model that is subtly wrong on a screen becomes obviously wrong when someone stands in it at their own height.
Choosing a Route
If your audience uses Quest and you want a walkthrough with no assembly, capture in Matterport and use their VR Mode.
If you work in the Apple ecosystem and want to stand inside your own captures, Vision Pro lets you do that directly from a scan you already have.
If you need the model itself for a game engine, a client’s platform, or a design pipeline, export a GLB and build the experience where you need it.
And if your audience will be on a laptop, a shared 3D model or a 3D floor plan covers it without the export work. Most clients are more comfortable opening a browser link than putting on a headset, so that route is often the more useful one anyway.
Need to show someone the space?
FAQ
How do I create a 3D floor plan for a VR walkthrough?
Scan the space, then choose a route to the headset. Matterport includes a WebXR VR Mode that runs on Quest with no app install. Polycam captures open natively on Apple Vision Pro. For other headsets, export a mesh as a GLB and load it into a VR platform or game engine.
Can I view a Polycam scan in VR?
On Apple Vision Pro, you can. The app runs natively on visionOS 1.1 or later, and captures made on an iPhone or iPad open in the headset. That is viewing a capture rather than an authored VR tour, and there is no native Quest app, so Quest requires exporting a model to another platform.
What file format do I need for VR?
GLB in most cases, since it holds geometry and textures in one file and most VR platforms accept it. GLTF is the same format in text form, so a GLTF export often arrives as a .glb file. Check what your host platform accepts before exporting.
Why does my scan run badly in a headset?
The cause is almost always polygon count. VR renders a separate image for each eye at a high refresh rate, so a raw capture mesh that behaves on a laptop can stutter in a headset. Decimate the mesh and crop to the area you need before exporting.
Do I need a headset to show a client a space?
No. A browser-based 3D model or a 3D floor plan communicates a layout perfectly well and asks nothing of the client beyond opening a link. VR earns its place when the audience already owns the hardware and standing inside the space tells them something a screen cannot.
What should I capture differently for VR?
Ceilings and floors need more attention than they get on a screen. Viewers look up and down in a headset in a way they rarely do on a screen, so gaps overhead or underfoot become obvious. Reflective surfaces also matter more, since phantom geometry is disorienting at full scale.
