The JSON Schema Standard

About this paper

JSON Schema is the most widely deployed API standard in the world, and hardly anyone thinks of themselves as using it. It does not arrive as a file — it arrives inside OpenAPI, inside AsyncAPI payloads, and now inside every agent tool definition. So it was measured the only way it can be measured: by opening 14,195 published API contracts and counting what is inside them.

9,439,146 schema objects. The largest measurement in this research line by two orders of magnitude. And the distribution is the story: eight keywords appear in the majority of documents, and every one of them existed before 2013. if/then/else, draft-07’s headline feature from 2017, appears 134 times across 9.4 million schema objects. unevaluatedProperties, the flagship of 2020-12, appears 230. Half the corpus sits nominally on the newest dialect and writes like it is on the oldest. format appears 887,362 times and is, by default, an annotation that no validator is obliged to check — which an agent will not know. The full report has the keyword-by-keyword census, the format problem, the dialect fossil record, the body compared head-to-head with the OpenAPI Initiative and AsyncAPI, and why the one standard here with no formal ratification is the only one that ships a conformance test suite.

What's inside

1. Executive summary — 9.4 million schema objects
2. What JSON Schema is, and where it actually lives
3. The body, and the ratification question
4. Nine million schemas — what the industry actually writes
5. The format problem — 887,362 uses of an annotation
6. The dialect story — the 3.0/3.1 split is a JSON Schema story
7. Reuse — the one thing the industry does well
8. The regulatory layer, the agentic turn, and where this goes
9. The scoring caveat, and the limits of this measurement
10. What would have to be true
11. The investable thesis
12. Where this is going

What you get for $500.00