The State of Headless APIs

About this paper

Headless is being talked about again as though it were one market. It is not. It is a delivery architecture that has been sold into three unrelated markets with three unrelated buyers — content, commerce, and rendering — that share a word and almost nothing else. What they do share is an unusually testable promise. When a vendor goes headless it gives up the defence every other software company gets to use: there is no UI to hide behind, no “the API is a secondary integration path.” If you sell headless, the contract is the product.

So I scored one hundred and eight of them the way an integrator or an agent would read them. The cohort averages 44.1 — better than most of the economy, with six genuine Exemplars. Then you split it by tier and the story inverts: the monoliths that headless vendors were funded to disrupt out-publish them, the headless CMS market finishes last of four, and WordPress beats thirty-three of the thirty-four headless CMS companies in the study. Not one provider in one hundred and eight signals idempotency, in a sector selling checkout as an API. The browser tier pitching itself to AI agents has published nothing an agent can use. The full report maps the commodity resource taxonomy, the facet-by-facet inversion, the twelve agent dimensions, the blueprint of what nobody satisfies, and discloses in full where these scores rest on artifacts API Evangelist generated rather than providers published.

What's inside

1. Executive summary — one word, three markets, one falsifiable promise
2. The state of the sector — one word, three markets
3. The resource taxonomy — a total commodity with no differentiation
4. The composite Kin Score — and the inversion
5. Agent-readiness — the frontier nobody has claimed
6. Provenance — what is theirs, and what is mine
7. The blueprint
8. Security, auth and scopes
9. Provider-by-provider intelligence
10. The investable thesis
11. Where this is going

What you get for $500.00