BUILDBOOK // MEMORY FOR THINGS I MAKE
Stop solving the same problem twice.
BuildBook grew out of a simple frustration: every prototype teaches you something, but that lesson is useless if the next version forgets it.
01 // THE PROBLEM
A folder full of files is not project memory.
CAD, STL and 3MF files tell you what existed. They do not reliably tell you why a dimension changed, which tolerance actually printed, what failed last Tuesday or which interface must never be “improved” back into an old bug.
02 // THE SCHEMA
Give the project a stable shape.
The project schema became one of the first foundations: known fields, units, versions, outputs and relationships instead of free-form notes scattered across folders. It gives both software and AI something concrete to read instead of asking them to infer the project every time.
03 // THE BUILDBOOK
Add the reasons, evidence and scars.
A BuildBook records design intent, validated dimensions, failed approaches, print results, decisions and release state. The schema says what the project is; the BuildBook preserves more of why it became that way.
04 // AI
Less guessing. Better continuity.
A schema does not make an AI deterministic, and a BuildBook does not make it infallible. What they do is reduce the amount of rediscovery and guessing. Instead of “make another version,” the system can be given the canonical dimensions, constraints, previous failures and expected deliverables before it starts.
05 // WHERE IT LED
Parametric designs became the seed for this website.
Once a design had explicit parameters and repeatable outputs, the obvious next question was: why should a user need my CAD environment at all? The Soccer Bag Tag customiser is the first public proof of that path — structured project knowledge behind a simple web interface that returns finished printable files.
The direction
BuildBook is evolving from personal project discipline toward a GUI-driven system that can work with local AI, cloud AI or a hybrid of both. The goal is not “AI designs everything.” The goal is preserving enough engineering context that tools — human or AI — can make the next decision with the previous evidence still attached.