How to 3D Scan with Your Phone or iPhone (2026 Guide)

What 3D scanning with a phone means

Learning how to 3D scan with your phone is mostly about taking the right pictures. The camera you already carry can capture enough information to rebuild an object, a room or an outdoor scene in 3D; no laser rig or turntable required.

The idea is simple. As you move around a still subject, the camera records it from many overlapping viewpoints. Reconstruction software finds the same visual details in different frames, works out where the phone was for each one, and then builds a 3D result: a point cloud, a textured polygon mesh, or a Gaussian splat.

Software cannot rescue poor footage. Most of the quality of a scan is decided while you shoot, so lighting, overlap and choice of subject matter more than any processing setting.

Photos or video?

Your first decision is whether to take individual photos or record a continuous video.

Photos give the sharpest frames. You get the camera's full resolution, no video compression, and more control over focus and exposure. Photogrammetry, which builds textured meshes, benefits from that sharpness. The trade-off is time: you stop, steady the phone and shoot dozens of angles.

Video is faster and gives even coverage. A short, slow walk around a subject produces hundreds of frames for the software to choose from. Video works especially well for Gaussian splatting, which likes dense, continuous camera paths. Walk slowly and shoot in good light so frames stay sharp, because motion blur is the most common reason video scans fail.

One practical detail for iPhone users: with the camera format set to High Efficiency, photos are saved as HEIF/HEIC. If your desktop software only accepts JPG or PNG, switch Settings > Camera > Formats to Most Compatible, which captures JPEG photos and H.264 video, or convert the photos before importing.

Lighting, surfaces and why scans fail

Reconstruction relies on finding the same features from different angles. Anything that makes a surface look different from each viewpoint, or look the same everywhere, causes trouble.

Lighting:

• Soft, even light is best. An overcast sky or light bounced off walls and ceilings gives gentle shadows.
• Avoid hard light. Direct sun, a flash or a bare bulb casts sharp shadows that can confuse the software.
• Keep it consistent. Do not change camera settings mid-capture, and avoid flickering or mixed-color lights.

Materials:

• Easy: matte, textured surfaces such as stone, raw wood, fabric, matte plastic and painted ceramics.
• Hard: reflective finishes such as chrome, glossy paint, glazed tiles and mirrors. The reflections move as you move, so the software cannot find stable points. A removable matte scanning spray can help on objects you are allowed to treat.
• Very hard: clear glass and acrylic. They reflect part of the light and bend the rest, so both their surface and what you see through them change with every step. Expect missing or smeared areas.
• Plain single-color or very smooth surfaces offer too few features to match.
• Movement: the subject must stay still. Pets, people, leaves in the wind and shifting fabric cause tears, blur or ghosting.

Apple's Object Capture guidance gives the same advice: diffused light, consistent settings, and avoiding reflective, transparent, translucent or very thin objects.

How to 3D scan with iPhone or Android: orbit paths

Whether you use an iPhone or an Android phone, move in planned loops rather than wandering. For a single object, three orbits usually cover it well:

• Low orbit: hold the phone below the middle of the object, tilted slightly up, and walk a full circle to capture the lower sides and any overhangs.
• Eye-level orbit: hold the phone level with the middle of the object and walk a second full circle.
• High orbit: hold the phone above the object, tilted down, and walk a third circle for the top surfaces.

Keep plenty of overlap. Apple recommends that neighboring photos overlap by 70% or more and warns that below 50% the reconstruction may fail. Before you start, lock exposure and focus: on iPhone, long-press the camera view until AE/AF Lock appears, so brightness and focus do not jump between shots.

Scanning rooms and outdoor scenes

Rooms and outdoor spaces need a different pattern from a tabletop object.

• Walk the edges. Move along the walls with the camera pointed into the room and toward the opposite walls. Do not stand in the middle and spin; the software needs the camera to change position to judge depth.
• Overlap wall passes. Sweep each wall in overlapping bands so corners and door frames appear in several views.
• Cover floor and ceiling. Include some angled views of the floor and ceiling from different spots.

Some iPhone Pro and iPad Pro models include a LiDAR Scanner, which measures depth directly. Apps use it for quick room layouts and scale, and Apple's guided Object Capture sample requires it. LiDAR depth is much coarser than camera detail, though, so for realistic textures and splats the camera images still do most of the work.

Processing on the phone or in the cloud

Once you have your capture, something has to turn it into 3D. Phone apps take two broad approaches.

• On the device: Niantic describes Scaniverse as creating Gaussian splats and meshes processed on the device, which gives quick feedback on site so you can reshoot missed angles. Its newer plans use monthly processing credits, while the classic Scaniverse experience keeps free on-device processing. Phones have less memory and cooling than a desktop, which limits how large a capture can be.
• In the cloud: KIRI Engine's pricing page explains that it pays for servers to process users' scans. Epic's launch announcement for RealityScan Mobile says the app assembles models using cloud processing. Polycam's web tool creates Gaussian splats from photos or a video you upload.

Cloud processing can handle heavy jobs, but it means uploading your files, sometimes waiting in a queue, and often paying a subscription for higher limits. If your scans show private homes, client spaces or unreleased products, think about whether you are comfortable sending them to a third-party server.

Processing locally on a desktop

For the most control, many people capture on the phone and process on a computer. Local processing has clear advantages:

• Privacy: your photos, videos and results stay on your own drive.
• More headroom: a desktop or laptop usually has more memory, a stronger GPU and better cooling than a phone, so it can handle larger photo sets, within the limits of your hardware.
• Free iteration: with local software you can reprocess with different settings as often as you like, without spending credits.

Getting files across is easy. AirDrop is the simplest route between an iPhone and a Mac, though large videos can take a while; a cable or a shared folder works too. On Android or Windows, copy the files over USB. Transfer the original files rather than versions shrunk by a messaging app.

Turning phone captures into Gaussian splats with SplatLoom

If you want to turn phone captures into Gaussian splats on your own computer, without a terminal or a subscription, SplatLoom is a desktop app for Mac (Apple Silicon, macOS 15 or later) and Windows 10/11 x64. It turns overlapping JPG or PNG photos, or an MP4 or MOV video, into a 3D Gaussian splatting scene entirely on your computer, with no upload, no account and no cloud processing.

Pick photos or a video, click Generate, then view and clean up the scene with box or sphere crop, rectangle select, one-click spike cleanup and undo/redo. You can export a Gaussian PLY, a PNG still, a 12-second 1080p MP4 orbit video, or a self-contained offline HTML viewer. It bundles COLMAP for camera positions, Brush for splat training on your local GPU and PlayCanvas for the viewer. In our test, a 13 to 20 second smartphone video was processed end to end in about 9 to 14 minutes on an Apple M5 Pro MacBook. Typical photo sets are 30 to 150 overlapping photos.

SplatLoom is a US$19.90 one-time purchase covering the Mac and Windows installers, with no subscription or credits, and a 7-day refund if it does not run on your computer. It does not export meshes, it is not a CAD or measurement tool, results depend on capture coverage, and its Windows GPU support is newer than its Mac support.