October 2025. A golfer in Eugene, Oregon is practicing in his garage with orange foam balls. His launch monitor can't detect them. He considers spray-painting them white.
Five months later, he built a launch monitor that works with colored balls and it's more accurate than devices costing five times as much.
The Garage
An iPhone propped on a takeout container. A hitting mat. A net. Orange foam balls scattered across the floor. Grandfather's vintage Wilson Black Heather irons from the 1970s, roughly two clubs weaker than anything made today. Every measurement during development was validated against harder-to-measure equipment. If it works with those clubs, it works with everything.
The person swinging them was running a restaurant grand opening, pulling 50-hour weeks on-site, and building a launch monitor in the hours around it. Early mornings, late nights, garage sessions between shifts. When his wife and son went up to Seattle over the holidays, he had extra hours to hit balls and test code before joining them.
This wasn't a sabbatical. It was a guy with a demanding full-time job who couldn't stop thinking about the problem. No venture capital, no lab, no team. A garage in Eugene with a net and a phone.
What Didn't Work
Approach 1: Optical Flow
Track the ball frame-to-frame using motion analysis. The problem: at 130 mph, a golf ball moves nearly 10 inches per frame. By the third frame, it's two feet away. The algorithm assumes small movements. Golf balls don't move small. Abandoned after six days.
Approach 2: Custom Dot-Marked Balls
Thirty days of engineering. 3,224 messages in a single conversation thread tuning GPU-accelerated corner detection. Custom balls with dot patterns for spin tracking. Rolling-shutter geometry for 3D pose estimation.
It worked in theory. In practice, the dots were too small to detect under garage lighting, the patent landscape for marked balls was hostile, and building custom balls made us a hardware company. That was the opposite of the thesis.
We killed 30 days of work because the honest answer was that it wasn't reliable enough to ship.
Better to measure three things accurately than six things poorly.
Then the pivot: what if we stopped trying to detect dots on a ball and started detecting the ball itself? What if the color, the orange that every other product ignored, was the signal?
The Breakthrough
Video cameras don't see in RGB. They capture in YCbCr, a format that separates brightness from color. The Cr channel measures red-orange saturation. An orange golf ball lights up the Cr channel like a beacon. The club, the mat, the net: none of them register. The color science suggests matte red and pink should land in the same high-Cr region, and we plan to test them next.
We could have made it work with any ball. It would have been less accurate. We chose accurate. Right now that means orange.
February 28, 2026. First validated speed measurement. 39.6 mph on a chip shot. An iPhone camera measured a golf ball's speed using nothing but light and color. It was rough. It was one shot. But it was real.
The Build
Orange balls aren’t detected. The problem is identified.
“Build an iPhone launch monitor.” Day one.
First code compiles. Four days from idea to running app.
30 days of engineering on a marked-ball system. Killed it.
The pivot to color detection. Orange balls become the answer.
Feb 28, 2026
39.6 mph. A chip shot. An iPhone just became a launch monitor.
Brand named. Acuvis Golf — “precision sight.”
30-shot volume test. 100% measurement rate.
Launch angle working. Two of three measurements in one day.
Carry distance announced in yards. The app speaks golf.
Volume test gate cleared. Calibration unblocked.
Four days from idea to running code. Five months from running code to validated measurements.
How It Was Built
Acuvis was built by one person with AI as a development partner. The architecture, the physics, every line of code: designed in conversation, tested in the garage, validated against the real thing. 221 conversations. Four complete engineering approaches. Three killed, one shipped. No outside funding. No team of 50. The technology was already in your pocket.
What We Believe
“No measurement is better than a wrong measurement.”
The app tells you when it’s uncertain. It never inflates a number.
“You shouldn’t need a subscription to know how fast you swing.”
Core measurement is a one-time purchase. Period.
“The phone in your pocket is the only instrument you need.”
No hardware, no dongles, no sensors. Software only.
“Practice should have data, not just feel.”
Every golfer deserves the feedback that used to cost $600+.
“We chose accurate.”
Colored balls are a deliberate choice. Orange is what we’ve validated. Red and pink should work based on the same color science, and we’ll test them next. The constraint that makes the measurement better.
The Founder
Cory Meza is based in Eugene, Oregon. He started Acuvis because he wanted launch monitor data for his practice sessions and couldn't justify $600 for a device he'd use in his garage. He built the first prototype while running a restaurant grand opening, pulling 50-hour weeks at the restaurant and fitting engineering sessions into the gaps. His test equipment: an iPhone on a takeout container, a hitting net, and his grandfather's Wilson irons from the 1970s. Acuvis is his first company. He still practices in the same garage.
Acuvis Golf is a product of Lyra Motion Technology. Eugene, Oregon.