Do You Need LiDAR for an AR Furniture App? Simple Guide

You are scoping an AR furniture app, and someone on the call says the words that stall the whole project: "should it require LiDAR?" It sounds like a hardware detail. It is really a decision about how many people can use your app at all. Get it wrong in the wrong direction and you either ship a weaker experience than you could have, or you lock out most of your buyers on day one.

Here is the short version before the detail. You almost certainly do not need to require LiDAR. You may well want to use it. Those are two different choices, and this guide keeps them apart.

What is LiDAR, in one plain sentence?

LiDAR is a small scanner that fires invisible laser pulses and times how long each one takes to bounce back, turning those timings into a depth map of whatever is in front of the camera. That is the whole idea. Instead of guessing distance from a flat photo, the phone measures it directly, many times a second, across the scene.

That depth map is what feeds the useful AR features. Once the phone knows the real distance to the floor, the wall and the coffee table, it can drop a virtual sofa in the right place, hide it behind real objects, and tell you roughly how wide the gap by the window is.

Which devices actually have LiDAR?

LiDAR is on the Pro and Pro Max iPhones from iPhone 12 Pro onward, and on iPad Pro from 2020. Standard iPhones, the SE line and every mainstream Android phone do not have it. That single fact shapes the whole decision, so it is worth sitting with.

If you require LiDAR, your app runs only on Pro iPhones and iPad Pro. That is a real audience, and often a higher-spending one, but it is a fraction of the phones your customers carry. A furniture retailer selling to the general public cannot afford to greet most visitors with "your phone is not supported". This is the device-support reality the glossier articles skip, and it is the reason the rest of this guide leans the way it does. If you are also choosing platforms, our comparison of how ARKit and ARCore compare covers device reach in more depth.

What does LiDAR add to an AR furniture app?

LiDAR improves four things: it locks onto floors and walls almost instantly, it powers clean object occlusion, it reads depth directly for measurement, and it can mesh a whole room in real time. None of these are new features you cannot get without it. They are the same features, done faster and more reliably.

Faster plane detection

Plane detection is how an AR app finds the flat surfaces, floors, walls, table tops, that it can anchor an object to. Without LiDAR, the phone builds these planes from camera motion and visual detail over a second or two of you moving the phone around. With LiDAR, the depth map gives the surfaces almost at once, even a bare white wall that a camera-only app would struggle to see.

Occlusion that looks real

Object occlusion means a virtual item correctly hides behind a real one. Place a virtual armchair partly behind a real sofa and, with LiDAR, the armchair is neatly cut off where the sofa is closer to you. Because LiDAR knows the true depth of everything in the room, these edges are clean. Without it you can still do people occlusion, where the app hides objects behind a person, and a rougher depth based occlusion, but the cut is less precise around furniture. Apple exposes this real-world depth through its scene depth data on LiDAR devices (ARKit sceneDepth).

Depth and measurement

A tape-measure feature is common in furniture apps: will this cabinet fit in that alcove? LiDAR reads distance from timed light, so it holds up in low light and on plain surfaces. A phone without LiDAR estimates depth from motion, which is genuinely good but wants texture, light and a moment of movement. If exact fit is central to your product, start from LiDAR and fall back gracefully. We wrote a full guide on how to get accurate AR measurements that goes into the calibration side.

Room meshing and RoomPlan

LiDAR can build a live 3D mesh of a room, a job Apple calls scene reconstruction (ARKit scene reconstruction). Apple builds on this with RoomPlan, a Swift API that, in Apple's words, "utilizes the camera and LiDAR Scanner on iPhone and iPad to create a 3D floor plan of a room" (Apple RoomPlan). If your app plans whole rooms rather than placing single items, this is powerful, and it needs LiDAR. RoomPlan exports the room as USD or USDZ, so it fits the same file pipeline as your product models. We break RoomPlan down in our guide to Apple RoomPlan, and cover the formats in USDZ and glTF explained.

What still works without LiDAR?

Almost everything a shopper needs. On a phone with no depth sensor, ARKit and ARCore still detect surfaces and place furniture. ARCore finds planes by tracking "visually distinct features in the captured camera image called feature points" and looking for "clusters of feature points that appear to lie on common horizontal or vertical surfaces, like tables or walls" (ARCore fundamentals). That is the core of AR furniture placement, and it does not need special hardware.

Depth is available too. ARCore's Depth API "uses a depth-from-motion algorithm to create depth images" and works on ordinary phones, merging a hardware sensor "such as a time-of-flight (ToF) sensor" only when the device happens to have one (ARCore Depth API). So depth sensing AR is not a LiDAR-only club. The table below sets the two paths side by side.

Read the table as a story about trade-offs, not winners. The LiDAR column is better on speed, occlusion and low light. The no-LiDAR column wins the thing that pays the bills for a consumer app: it runs on almost every phone your customer already owns.

LiDAR vs no LiDAR: how to decide

Decide by your use case, not by the spec sheet. Require LiDAR only if whole-room scanning, hard occlusion or exact measurement are core to the app and your buyers already carry Pro iPhones. For a normal product catalogue, support every AR phone and switch the LiDAR features on automatically when the device has the scanner. The checklist below is the rule we use.

In code this is a simple pattern. At runtime the app asks the device whether it supports scene reconstruction or scene depth. If yes, it turns on occlusion, faster meshing and the tighter measure. If no, it runs the camera-only path. The shopper never sees a hardware error, and you never write two apps. This is progressive enhancement, the same idea the web has used for years, applied to AR.

Where LiDAR and depth sensing fail

LiDAR is not magic, and knowing its failure modes keeps you from promising something the hardware cannot deliver. Laser depth breaks down on a predictable set of surfaces, and camera-only depth has its own weak spots. Design around both.

• Mirrors and glass. Laser pulses pass through or scatter, so a glass table or a mirrored wardrobe confuses LiDAR. The mesh gets holes or false surfaces exactly where a furniture shopper is likely to be standing.
• Shiny and dark surfaces. Polished metal reflects the beam away, and very dark matte surfaces absorb it, so both read as unreliable depth. A high-gloss floor is a classic problem case.
• Very bright or very dark rooms. Strong sunlight adds noise to the return signal, and near-total darkness leaves the camera with nothing to track alongside the depth.
• Thin and fine shapes. Chair legs, lamp stems and railings are thin enough to slip between depth samples, so occlusion around them looks ragged.
• Range. LiDAR on these devices reads a few meters, not a whole warehouse. It suits rooms, which is fine for furniture, but do not sell it as a site-survey tool.

Camera-only depth fails differently. ARCore itself warns that "flat surfaces without texture, such as a white wall, may not be detected properly" (ARCore fundamentals), because it has no feature points to lock onto. A blank wall, a plain carpet or a dim room all slow it down. This is the honest picture: LiDAR removes some of these problems and adds a few of its own.

When requiring LiDAR is not worth it

Requiring LiDAR is not worth it when your goal is reach, which is most of the time for a retail app. The moment you make LiDAR mandatory, you cut your audience to Pro iPhone and iPad Pro owners and lose every Android user and every standard iPhone user in one line of code. For a brand that wants shoppers to try a sofa at home, that is the wrong trade.

There is also a build-cost angle. Supporting both paths, camera-only plus a LiDAR upgrade, is the standard approach and does not double the budget, because both sit on the same ARKit or ARCore foundation. A well-scoped AR MVP at appico starts from $10,000, a full mobile app from $12,000, and larger multi-platform builds run up to about $150,000 depending on the catalogue size and the 3D pipeline behind it. Building LiDAR-only does not save money; it just narrows who can pay you. If you are early, our guide on turning an idea into an app helps you scope the first version, and you can sketch a figure with our app cost calculator.

The other case where LiDAR is beside the point: the real cost driver in an AR furniture app is usually the 3D model catalogue, not the depth hardware. If your models are heavy, badly scaled or missing, no sensor will save the experience. We saw the same pattern in retail try-on work, which we wrote up in the cost and time to develop a virtual try-on app. Spend the budget where the app actually stands or falls, and reach for custom AR app development only once the scope is clear.

Our take

Treat LiDAR as a feature you switch on, never a gate you make people pass. We build AR furniture apps to run on every supported phone first, then light up the faster locking, occlusion and room scanning automatically when the device is a Pro iPhone or iPad Pro. The shopper with a three-year-old Android still gets to place the sofa. The shopper with a new Pro iPhone gets the sharper version. Nobody is turned away.

If you are weighing this decision for a real product, we are happy to talk it through against your catalogue and your audience. You can see how we approach AR and mobile app builds and bring the messy questions. The hardware is the easy part; the reach, the models and the honest scope are where the project is won.

Frequently asked questions