Sefism is free for learners. We are here to help you grow, one step at a time.Join for free
Sefism.
Games & XR

WebXR 3D and AR Product Viewer Pipeline for Furniture Retailers

A pipeline and embeddable viewer that lets small furniture and home decor sellers turn product 3D models into fast web 3D views and true-scale AR placement on phones, with automatic optimisation and simple analytics. An intermediate final year project joining 3D graphics, WebXR and ecommerce for a real retail market.

Topics
Difficulty

Intermediate

A small team, or one strong student willing to learn something new.
Effort

2 semesters · 3 students

Deliverables

7 to ship

3 optional extras

Suggested stack

three.js or model-viewerglTF TransformBlender (Python scripting)Node.jsNext.jsCloudflare R2 or S3
A suggestion, not a requirement. Swap anything for what you already know.

What you should ship

  • An upload portal where a seller adds a model (glTF, FBX or OBJ) with real dimensions and gets back an optimised GLB and a USDZ for iOS.
  • An optimisation pipeline: mesh simplification, Draco or meshopt compression, texture resizing and KTX2 conversion, with a report of size before and after.
  • An embeddable viewer script for any store page (Shopify, WooCommerce or a custom site) with orbit, zoom, material variants and a "view in your room" button.
  • True-scale AR placement: WebXR with hit testing on Android Chrome, and Quick Look with the USDZ on iOS, both checked against a tape measure.
  • A visual check that renders the original and optimised model side by side and computes an image difference score, so optimisation never ruins a product.
  • Seller analytics: views, rotations, AR launches and time in viewer per product.
  • Ten real products from at least one local seller processed end to end and live on their site or a demo store.

If you have time left

  • Generating a rough model from phone photos using an open photogrammetry tool, with a manual cleanup step.
  • A room planner that places several products together in AR.
  • A/B test of product pages with and without the viewer on add-to-cart rate.

The problem: will this sofa fit?

Furniture is one of the hardest things to sell online. Customers want to know the size, the colour under their own light and whether it fits the corner of their lounge. Returns are expensive because the item is bulky. Large retailers answer this with 3D viewers and AR placement. Small furniture makers and home decor shops, who sell through their own site, Instagram or a marketplace, rarely have the skills or budget to do the same.

Even when a seller has 3D models from their manufacturer or a designer, those files are usually far too heavy for a phone: millions of polygons and 8K textures, built for renders rather than the web.

What already exists and the gap

  • Shopify supports 3D models and AR on product pages using GLB and USDZ files, but the seller must supply web-ready models.
  • Google's model-viewer is an open source web component that displays 3D models and launches AR on supported phones.
  • Threekit and VNTANA offer enterprise 3D product visualisation and optimisation platforms with pricing aimed at larger brands.
  • IKEA has built its own AR placement into its app, which is not available to anyone else.

The gap is the middle: an affordable pipeline that takes whatever model a small seller has, makes it light and correctly scaled, checks it still looks right, and gives them one line of code for any store. WebXR support also differs by platform (hit testing works in Chrome on Android, while iOS relies on Quick Look with USDZ), so the product has to handle both paths.

What you build

Sellers upload a model and enter real dimensions. The pipeline normalises units and orientation, simplifies the mesh, compresses geometry and textures, bakes material variants and produces GLB and USDZ outputs, with a size and quality report. Shoppers see an embedded viewer on the product page, rotate it, switch fabrics and tap "view in your room" to place it at true scale. The dashboard shows sellers which products get the most engagement.

The technically interesting part

Optimisation without visible damage is the core problem. You must decide how far to simplify each mesh and shrink each texture to meet a mobile budget, then prove the result still looks like the product. Build an automated check: render the original and optimised models from fixed camera angles with the same lighting and compute an image difference metric, rejecting results above a threshold. Combine that with correct real-world scale, which is a common bug in AR (a chair that appears the size of a house), and cross-platform AR launch.

Show evaluators a chart of file size against visual difference score across your ten products, then a live AR placement measured with a tape.

How to evaluate it

  • Size and speed: median GLB size and time to first render on a mid-range Android phone on 4G, before and after optimisation.
  • Visual fidelity: image difference score per product and a blind comparison by 10 people shown original and optimised renders.
  • Scale accuracy: difference between AR size and tape-measured size for 5 products.
  • Seller effort: time for a non-technical seller to go from model to live viewer.
  • Shopper feedback: a short test with at least 10 people on whether AR changed their confidence about size.

Suggested plan

  1. Weeks 1 to 4: find a seller partner, collect models, study glTF, Draco and KTX2.
  2. Weeks 5 to 9: pipeline CLI with glTF Transform and a Blender script for format conversion.
  3. Weeks 10 to 14: embeddable viewer with variants; Android WebXR placement.
  4. Semester 2, weeks 1 to 5: iOS USDZ path, scale checks, visual difference checker.
  5. Weeks 6 to 9: seller portal, analytics and embed code.
  6. Weeks 10 to 14: process ten products, run evaluations, report and demo.

Path to a product

The first customer is a local furniture maker or home decor brand with an online store. Charge per product processed plus a monthly hosting and analytics fee, or offer a Shopify app with a per-store subscription. The same pipeline serves other bulky categories: appliances, bicycles, lighting. Later, offer model creation as a service using freelance 3D artists, with your pipeline guaranteeing quality.

Scope warning

Do not build a 3D modelling tool, a full room planner or your own AR tracking. Do not promise photogrammetry as the main input; it is a stretch goal. Pipeline, viewer, AR at correct scale and honest measurements are the project.

Learn what you need on Sefism

Get help with this idea

The risk here is spending months on 3D art; the pipeline and the measurements are what earn the grade. Bring your proposal and we can agree the scope, pick a stack your team can finish with, and design an evaluation your panel will respect. You build it; we help you make the calls.

New here? Create a free account first.

Ideas and guidance, not finished projects

These are project ideas and scoping guidance, published free for students to use as a starting point. I do not build, write, or sell final-year projects, and I do not complete coursework for anyone. Take an idea, make it yours, and build it.

Not sure this one fits you?

I help students pick an idea that matches their skills and their deadline, then scope it down until it is finishable. Guidance only, never done for you.

Discuss this idea
Work with me

Stuck on something specific?

Writing only gets you so far. If you want an answer to your situation rather than the general case, book a session and we will work through it together. Sessions are free for approved Sefism members, and a few slots open each week.

Book a free sessionSee all services

Follow along

New writing, resources and project ideas land here first.