Personalized Software

Producing software used to be expensive. A non-trivial app meant paying a programmer for months of work, or a whole team, and maintenance was not cheap either. That is still true in high-stakes environments. In a lot of other places it is not. It is now feasible to build software for a very small company, tailored to what that company actually does, and in some situations it makes sense to build it for a single user.

The interesting question is not whether this is possible. It is which things were never built because the maths did not work, and now do.

A home wall is a market of one

A spraywall in a garage is used by its owner and maybe six friends. No app was ever going to be written for that wall, so home wall climbers use apps written for everyone: pay per wall by hold count, work around the grading system on offer, accept that the wall shape or the angle you built is not one the app models.

None of that is unreasonable of the app makers. Serving everyone is what makes the economics work for them. But the gap between "an app for all spraywalls" and "an app for my spraywall" used to be a few months of somebody's salary, and now it is a weekend. A list of problems, a photo of the wall with holds tapped out, a logbook, a grade you and your friends argue about: that is a small application. It was never small enough to be worth building alone. It is now.

A gym is a market of one, too

Gyms are in the same position with more at stake. Which app your climbers use is usually decided by whoever sold you the board, and the software half of that purchase is not really yours. In March 2026 the Kilter Board's app disappeared overnight in a dispute between the board maker and the developer who ran it, and twelve years of logged ascents went with it. I wrote about what that looked like from the app side.

A gym that runs its own thing is not exposed to that. It also gets to make decisions no general app will make for it: a wall reset that keeps the old problems, a members' competition with your rules, a screen by the wall showing today's circuit, an export into whatever you already use for memberships. Each of those is a small feature that no vendor will build for one gym, and all of them together are still a smaller project than the board itself cost.

Where it does not pay off

Building your own is not free, and two things do not come with it. The first is the community: an established app arrives with thousands of problems and other people's beta, and an empty app you wrote yourself arrives with nothing. The second is that you now own the maintenance, forever, and "forever" outlasts the enthusiasm that started it.

So the honest version is narrower than the hype. Build your own when the thing you want is specific to you and the alternative is paying to be a rounding error in somebody else's product. Use what exists when what you actually want is the crowd.

The part that was hard

For a climbing wall, the piece that used to stop this was the wall itself. Turning a photo of a chaotic spraywall into holds you can tap is real computer vision work, and it is not what a gym owner or a home wall builder wants to spend their weekend on.

That is the part OwnClimb does as an API: photograph a wall, get scored hold polygons back, and build the small, specific thing you actually wanted on top of it.