JBCT is a deterministic, AI-friendly way to write backend Java: a small set of patterns and rules that reduce the space of valid choices until there’s essentially one good way to structure a use case. Code ends up reading like the business process it implements, whether a person wrote it or an AI assistant did, because the structure is the same either way.
The method rests on typed failures instead of exceptions, parse-don’t-validate at every boundary, functional composition (map/flatMap over Result, Option, Promise), and a fixed catalog of six structural patterns: Leaf, Sequencer, Fork-Join, Condition, Iteration, Aspects. Every decision in the method is checked against five criteria — mental overhead, business/technical ratio, design impact, reliability, complexity — not against taste.
The pitch is usually told as noise removal — technical detail pushed to the edges so the business flow stands out. The stronger half is what survives: the business facts live in the types. A return type states whether a step can fail; an Option parameter states that the domain allows absence; all() states that steps are independent; a sealed error type states the complete failure catalog. Hide the machinery, keep the meaning — the code executes and testifies at once, and the compiler keeps the testimony true.
The free course on this site walks the whole method with worked examples. The book, Java Backend Coding Technology, is the complete argument, on Leanpub. Pragmatica Core is the library the method is written against.