PLUed
Trains grocery cashiers on produce PLU codes with spaced repetition and a verified code library.
A studio that finishes things
The idea was never the barrier. Finishing is.
Arbolinea designs and builds apps, and makes objects in the workshop. We use AI tools all day, the way a woodworker uses a table saw: they do the heavy lifting, we decide what the thing should be. Then the part we care about most, getting it right the first time, or in far fewer tries.
Ideas above the treeline
Software
Apps
One app live in three stores, and three more on the way.
Crab Math Run
Math practice built to keep kids playing. A crab runs a pier over the sea, and a parrot drops a plank across each gap as the problems get answered. No ads, no accounts, nothing collected.
Timed Math Pro
Arithmetic drilling that rewards speed and accuracy. Built in 2012 to help kids get more confident with basic arithmetic, and in the stores for over a decade until it was pulled for being out of date. We are rebuilding it from the ground up.
Tetradec
A merge puzzle game built from regular polygons.
Physical
The workshop
Objects built by hand, with AI in the process. The saw is built. The cajon is still a model.
“Be Better” cajon
Designed, not yet cutA box drum, designed and engineered with AI. It will be cut and assembled by hand from Baltic birch.
Elmore Buck Saw
Shop-madeA bucksaw that folds into itself: two handles, a stretcher, a blade, and a cord tightened by a toggle. Folded, it fits a canvas pouch. We built it in the shop.
The practice
How we work
These tools are part of every day here. Using them well is a craft of its own, and it is the part we work hardest at.
-
Many tools, chosen per job
We keep several AI tools running and move between them depending on the work. One writes better prose, another is stronger at code, another is cheaper for a long research pass. Which one is best changes month to month, so we keep testing the new ones as they ship and we watch what each one costs.
-
We write our own skills
A skill is a set of reusable instructions that teaches an AI tool how we design, write and build, so we are not starting from scratch every time. Writing a good one is instructional design for a machine, and it takes the same care. We keep ours and refine them as the work teaches us something.
-
We design with them, not only code
The product and the interface get the same treatment as the code, and so does a physical piece. The cajon is the plainest example: AI helped design and engineer it, and the cutting and assembly will be done by hand. The same goes for the brand assets and for this page.
-
Built on how people learn
The studio comes out of nearly thirty years in training and data analysis, so when an app teaches something we use what works. PLUed drills produce codes with spaced repetition, because spaced practice is how codes like these stick. Crab Math Run never ends a run on a miss in its default mode, because a kid who is afraid of being wrong can stop trying.
-
Validation first, then release
Before we build, we find out whether the idea deserves the time: what already exists, where ours would differ, and what it would take. Nearly every project starts as a website, the cheapest way to find out if an idea is real. The few that need more become apps, and we take them properly through the App Store, Google Play and the Amazon Appstore.
What we hold
The Doctrine
we will not pretend
On the tool
Ideas come easy now. The machine can generate them, design the piece, and suggest a better one. We will not pretend it cannot. We also will not pretend we have its memory or its speed. What we have are wants: we know what we like, we know the future we want, and we build what makes us and others better. The machine has no wants. We do. So when we say AI, we mean exactly what an engineer means by a calculator, or a woodworker means by a table saw. A tool.
On the market
The tool is not free. There is a monthly bill, and a harder question under it: which one? Grok, Codex, Kimi, Claude, all of them, or whatever ships next week? Then which model, which mode, which plan? Is the new one better? Better at what: writing, code, price? The answers change month to month, sometimes day to day. We watch the landscape constantly and pick what fits the work. We will not pretend that choice is settled. It never stays settled.
On platforms
Nearly every project starts as a website: the cheapest, fastest way to find out if an idea is real. Most never need to be more than that. The few that do, we build as apps you install from the App Store, Google Play, or the Amazon Appstore, and we take them through those stores properly.
The studio
About
arbor + linea
the tree line
Arbolinea started with a problem - a math homework problem. In 2012 Clay Elmore built the company's first app, Timed Math, to help his kids be more confident with their basic arithmetic skills.
He spent nearly thirty years in training and data analysis, including learning programs at Nokia, where his work included launch calendars, the learning platforms, content reviews, testing with real users, translation, and dashboards that tracked who finished and what they thought of it. He has also taught people the basics of using AI tools and writing prompts. It is where the studio's instinct for teaching comes from, and it is the same instinct we use when we write a skill for an AI tool.
He is also a woodworker and a lifetime outdoorsman. The treeline in the name is a real place. It is the line where the trees stop and the view begins. He first saw that view on a summit attempt on Longs Peak, in Colorado. When you cannot see the forest for the trees, get above them. Every project since has started the same way. With a problem worth solving.
- Nearly thirty years in training and data analysis
- Learning programs at Nokia
- Woodworker
- Number theory research
Get in touch.
Questions about one of our apps, or about the work? Write to us.