TL;DR:
- AR furniture placement on phones (IKEA Place, Wayfair, Amazon) now works accurately enough for real purchase decisions thanks to LiDAR on modern iPhones and iPads.
- Professional interior designers use room-scanning apps (Polycam, RoomScan Pro) to generate accurate 3D floor plans as the basis for AR walkthroughs.
- Apple Vision Pro enables 1:1 scale room walkthroughs with full spatial context — the most compelling XR interior design experience available, though at a price point that limits consumer adoption.
Visualising how furniture will look in your home before buying it has been possible in theory since the first AR apps appeared. For years, the experience was disappointing — items floated incorrectly, scale was wrong, shadows didn’t behave realistically, and you’d buy the sofa anyway only to discover it blocked the radiator.
The experience in 2026 is genuinely different. The combination of LiDAR-equipped phones, improved understanding of room geometry, and better 3D asset quality from retailers has made AR placement accurate enough to make real decisions from. Here’s what you can actually do with it.
Phone-Based AR: Where It Works Well
IKEA Place remains the benchmark for furniture AR. Using ARKit (iPhone) or ARCore (Android), the app places IKEA products at 1:1 scale in your camera view using your phone’s spatial understanding of the floor and walls. On an iPhone 15 Pro or later (with LiDAR), it accurately prevents items from overlapping with real objects and adjusts to uneven floors.
What works well: placing large furniture items (sofas, wardrobes, dining tables) to check size and fit. Seeing whether that 3-seat sofa physically fits between the window and the TV unit is something most people misjudge from measurements alone — AR makes it immediately obvious.
What still has limits: fine-grained material and colour accuracy under different lighting conditions. A sofa that looks grey in the camera might read differently in warm evening light. Use AR for spatial decisions, not colour matching.
Wayfair’s View in Room and Amazon’s View in Your Room work similarly. The product catalogues are extensive — most major furniture listings have 3D models — and the placement quality is comparable to IKEA Place for large items. Both support saving room views with items placed, useful for comparing options later.
Houzz: Beyond AR placement, Houzz’s app includes a room-scanning feature that builds a rough 3D model of your space, which you can then furnish virtually from their catalogue. The model quality isn’t precise enough for professional work but is good enough for getting a rough impression of a full room layout.
LiDAR Makes the Difference
If you’re using AR for serious purchase decisions, the LiDAR sensor on iPhone Pro models (iPhone 12 Pro and later) and iPad Pro significantly improves accuracy. LiDAR builds a depth map of the room in real-time, letting the AR system understand precisely where walls, floors, furniture, and obstacles are — not just estimating from the camera image.
The practical effect:
- Furniture placement is more accurate (less drift as you move the phone)
- Items don’t “float” above the floor or sink into it
- The app understands the edge of a rug or the front of an existing sofa, so new items can be placed in relation to them correctly
- Room boundaries are detected faster
For complex layouts — rooms with alcoves, split levels, or a lot of existing furniture — LiDAR is meaningfully better than camera-only AR.
Room Scanning for Accurate Floor Plans
AR furniture placement is useful for quick checks. Professional planning — renovation, significant refurnishing, layout changes — needs an accurate floor plan first.
Polycam (iOS/Android) uses LiDAR and photogrammetry to build a textured 3D model of a room. You scan by slowly moving your phone around the space; the app stitches the depth data and photos into a 3D mesh you can export as USDZ, OBJ, FBX, or DWG. Output quality is high enough for renovation planning and is used regularly by interior designers as a site survey tool.
RoomScan Pro is faster to use but produces a schematic floor plan rather than a full 3D model — point at walls, ceiling, and floor for each room, and it generates accurate dimensions automatically. Exports to PDF, DXF, and directly to home planning software.
Canvas (iPad only, by Occipital) produces the most accurate room scans of any consumer app, using the iPad Pro’s LiDAR sensor and their proprietary scanning algorithm. The output is a dimensionally accurate 3D model that imports correctly into architectural software. Canvas is used by professional designers, real estate firms, and renovation contractors. A single scan exports as USDZ or STL, with a paid upgrade to DWG/DXF for professional software.
Designing in AR: Beyond Furniture Placement
Once you have an accurate room model, you can go further than placing individual items:
Full layout exploration: Apps like Planner 5D let you build out a complete room or home layout from your floor plan, furnish it, and then switch to AR mode to walk through it overlaid on your real space. The furniture library is extensive and you can customise finishes, colours, and dimensions.
Paint and wallpaper visualisation: Dulux Visualiser (UK) and Sherwin-Williams ColorSnap let you “paint” your real walls in AR — the camera view is processed to recolour the walls in your chosen colour with realistic shadow and texture. Handy for ruling out colour choices before committing.
Lighting planning: Some professional tools (Dialux Evo, Relux) allow you to import room models and simulate lighting conditions — where shadows will fall, how different light sources interact. This isn’t consumer AR but is used by interior designers for client presentations.
Apple Vision Pro: The Most Complete Experience
For anyone with access to an Apple Vision Pro, the interior design use case is one of the most compelling real-world applications. The combination of passthrough AR at high resolution, accurate 3D understanding of the room, and spatial audio creates an experience qualitatively different from phone AR.
IKEA’s spatial app for Vision Pro places furniture at 1:1 scale with accurate physics — sit in your actual chair and look across your room at the virtual sofa where it would actually be, at the correct height, correct width, correct distance. Rotate it, change the cover colour, walk around it. The experience of a well-executed Vision Pro furniture demo is genuinely impressive as a decision-making tool.
The limits are commercial: at £3,499 in the UK, Vision Pro is a product-category demonstration rather than something most people use for home planning. For professional designers, it’s beginning to appear as a client presentation tool — showing clients a finished room in 1:1 scale before committing to a renovation is a compelling differentiator.
Practical Tips for Getting Good Results
Light the room properly: AR tracking works better in good ambient light. Dim rooms confuse depth estimation and cause objects to drift.
Clear the floor: For furniture placement, remove clutter from the area where you’re placing virtual items. AR systems work better when they can cleanly identify the floor plane.
Measure first, AR second: Use AR to confirm and visualise measurements, not to replace them. Measure your space with a tape measure, check items fit in the measurements, then use AR to see how they look.
Use manufacturer apps where available: IKEA, Wayfair, and Loaf all maintain their own AR implementations with accurate models of their actual products. Third-party AR furniture apps often use lower-quality models with inaccurate dimensions.
Export your scan for future reference: Once you have an accurate Polycam or Canvas scan of a room, export and keep it. Future furniture searches become much easier when you can drop items into an accurate model of the space without re-scanning.
The technology has genuinely arrived for practical use. The main remaining gap is model coverage — many smaller retailers and specialist furniture makers don’t have 3D assets for their products. For the major furniture retailers, AR is now a useful part of the buying process rather than a novelty.