1600-SF23-Illuminator XR

Inspiration

We realized that lighting dramatically changes how a room feels — but choosing the right fixture is surprisingly difficult. Most people don’t know how bright a space should be, what style fits their layout, or how a fixture will actually look in their home. We wanted to solve this with something more intuitive than online shopping or interior design guesswork.

With the rise of mixed reality and AI, we saw an opportunity: What if your room could tell you exactly what lighting it needs? And what if you could preview that lighting instantly in your real space?

That idea became IlluminatorXR.

✨ What it does

IlluminatorXR is an AI-powered mixed reality lighting designer for Meta Quest.

It allows users to:

Scan their real room using passthrough MR
Detect furniture layout, ceiling geometry, and current brightness
Analyze the space and identify lighting gaps
Receive intelligent lighting recommendations
Preview light fixtures instantly in mixed reality
Adjust brightness and color temperature in real time
And for fun… activate Disco Mode to transform any room into a party

In short, it helps users see their space in a new light.

How we built it

🛠 How we built it

Built on Unity with Meta XR SDK, using Passthrough and Scene Understanding
Used Quest’s spatial mesh and plane detection to understand room geometry
Captured passthrough frames and sent them to a Roboflow AI detection service
Analyzed brightness and estimated color temperature directly from camera input
Designed a rule-based recommendation engine to suggest fixtures based on room layout and lighting needs
Implemented MR placement of 3D light fixtures with adjustable intensity and warmth
Added dynamic lighting effects, glow shaders, and a playful Disco Mode sequence
Designed a clean UI system to guide scanning, analysis, and selection

🧩 Challenges we ran into

Meta SDK passthrough relighting work differently in different version of the Meta SDK
Performance constraints on the Quest meant we needed lightweight models
Combining scene mesh and AI detection make scene extremely heavy and hard to move forward
Getting lighting previews in Unity to feel “real” with only approximated environmental data could be hard, need more shader work.

🏆 Accomplishments that we're proud of Illuminator XR

Built a working end-to-end MR pipeline in only around 2 days
Achieved real-time room scanning + AI analysis on the Quest
Designed a delightful MR lighting preview that feels surprisingly natural
Created Disco Mode — the demo moment everyone smiles at
Developed a clean, judge-friendly user flow from scanning → analysis → preview
Produced a polished UI that clearly communicates the vision
Training a custom lighting fixture detection model

Most of all, we built something that genuinely feels useful and magical.

📚 What we learned

Lighting perception is more emotional than technical — small changes have huge impact
AI models don’t need to be perfect to create meaningful recommendations
Good storytelling and user flow matter as much as the tech in a hackathon
MR UX requires minimalism, clarity, and constant anchoring
Quest passthrough and spatial mapping are incredibly powerful tools when paired with AI
Good storytelling and user flow matter as much as the tech in a hackathon

🚀 What's next for IlluminatorXR

Adding lux measurement + lighting health scoring
Detection more furnitures etc, kitchen table for cooking, working desk, bookshelf, dinning table to analyse and give specific suggestion
Partnering with lighting brands to allow one-click purchasing
Supporting multi-fixture planning and automatic lighting layouts
Bringing IlluminatorXR to phones and AR glasses for broader access

IlluminatorXR is just the beginning of smart, AI-driven interior design — and we’re ready to push it further.