Introducing Enhance

I built a digital twin of my university using cameras it already has — no new sensors, and it runs on a laptop. The school wasn't interested. I still think it's the coolest thing I've shipped this year.

Try it: enhance.lintware.com

_Watch on YouTube: Introducing Enhance_

Singapore's scarcest resource is land, and we waste a shocking amount of it every night. Walk past any campus or office block after hours: lights on, air conditioning running, rooms empty. We've spent twenty years putting sensors in our pockets, but our buildings are still dumb. Nobody is watching them.

The pitch: a digital twin with zero new sensors

The usual answer to "how many people are in this building" is to install sensors — PIRs, Wi-Fi tracking, badge readers. Expensive, invasive, and slow to roll out.

But the cameras are already there. Every corridor, every classroom, every carpark already has one pointed at it. The footage exists. We just weren't extracting anything from it.

So I built Enhance — a proof of concept that turns existing camera feeds into a live digital twin. I pointed it at my campus and put every room on a map: click any room and you get a live headcount, the floor, the building, the status.

What's in the demo

Campus map. Every building outlined, every room searchable. Satellite view, plus a 3D mode you can flip between solid and wireframe. A map of the whole campus that knows how many people are in each room right now.

Image analytics (enhance.lintware.com/vision). Feed it a single image and it reads the scene — what's happening, what's wrong, what materials are present. You can even prompt it: show it a room and ask "is there a bag?" and it'll tell you.

Image → 3D. Give it photos and it reconstructs a 3D model of the space. No LiDAR, no depth camera — just regular images.

The part that surprised me

Everyone assumes this is expensive. It's not — the whole thing runs on my laptop. The models are efficient enough that you don't need a GPU farm to do per-room analytics across a campus.

That's what makes the scale story interesting. If it runs on a laptop for one campus, the same approach grows to a whole city. Cameras scale cheaper than sensors ever will.

The honest part

I tried to get the school interested. Walked the demo around, explained the pitch — nobody took it up. Fair enough; universities move slowly, and an unsolicited demo is easy to wave off.

But the math doesn't change: the cameras are already mounted, the insights are sitting there unread, and the compute is cheap. Somebody is going to build this for real. I just wanted to be the one who showed it could run on a laptop first.