Who maintains the specifications this industry runs on

Every organization reading this depends on the specifications below. Almost none of them could name five people who maintain one.

This page is the list, generated from public GitHub contribution counts and nothing else — a login and a number of commits to a named repository, both re-checkable by anyone. No employers, no affiliations, no inferred anything. The script that builds it is in the repository, so you can run it yourself and get the same answer.

It is here to support one argument: the machine-readable substrate of the modern API industry is maintained by a countable number of named individuals, and the relationship between that number and the size of what depends on it is the actual problem. The funding argument is here.

Caveats, because a list like this can mislead

Commits are a terrible proxy for contribution. Specification work is argument, review, issue triage, editorial judgement, and turning up to meetings for years. A one-line commit merging a decision can represent six months of work by people whose names are nowhere near it.
The endpoint counts commits and caps at 100 entries per repository. Treat every contributor count as at least this many.
Repository choice shapes the answer. Several of these projects do their real work across several repositories; this counts the one where the specification or the core tool lives.
Absence means nothing. Plenty of the most consequential people in this space do not appear below, and a few of the most prolific committers would be the first to say so.

Read it as a rough census, not a leaderboard.

The specifications

AsyncAPI

asyncapi/spec — 70+ contributors

fmvilas 375 · asyncapi-bot 116 · derberg 46 · char0n 16 · MikeRalphson 12 · smoya 12 · SensibleWood 11 · bpedro 8 · jschabowsky 6 · magicmatatjahu 5 · Pakisan 5 · BlockLucas 4

JSON Schema

json-schema-org/json-schema-spec — 63+ contributors

handrews 753 · gregsdennis 418 · Relequestual 211 · awwright 192 · jdesrosiers 109 · fge 101 · geraintluff 88 · kriszyp 46 · nickl- 36 · karenetheridge 30 · dlax 26 · garycourt 16

The tooling

Spectral

stoplightio/spectral — 97+ contributors

StoplightSupport 243 · P0lip 100 · casserni 60 · nulltoken 47 · mnaumanali94 43 · tbarn 43 · XVincentX 27 · semantic-release-bot 26 · dependabot-support 21 · tekyu 18 · micryc 15 · heitortsergent 14

vacuum

daveshanley/vacuum — 47+ contributors

daveshanley 1454 · tx3stn 27 · calvin-lobo-hs 17 · rica-graca 16 · TheTeaCat 13 · scop 13 · Marat-Tim 7 · Karan-Palan 7 · volovikariel 7 · eli-l 5 · andreyyudin 4 · califlower 3

Redocly CLI

Redocly/redocly-cli — 98+ contributors

RomanHotsiy 378 · tatomyr 376 · DmitryAnansky 156 · knidarkness 106 · andriyl 84 · IgorKarpiuk 64 · lornajane 42 · AlbinaBlazhko17 41 · adamaltman 41 · jacobator 40 · AntonKozachuk 38 · vadyvas 33

SpecLynx ApiDOM

speclynx/apidom — 2+ contributors

char0n 210 · thanatosbot 1

Zally

zalando/zally — 59+ contributors

netme 526 · maxim-tschumak 439 · roxspring 358 · vadeg 168 · fmueller 159 · rbarilani 107 · zeitlinger 83 · mkulak 65 · superbenji 47 · thomasabraham 20 · akryvko 18 · bhaskarmelkani 12

What the shape of this tells you

Look at the distribution rather than the names. In almost every project on this page, a very small group carries most of the commits, and the tail falls off a cliff. That is not a criticism of anyone — it is the normal shape of specification work, and it is exactly the shape that makes a project fragile when one or two people change jobs.

Now put that next to what depends on it: national standards, procurement documents, hyperscaler products, and effectively every API programme in the industry.

The gap between those two facts is the thing nobody has solved. Not for this format, and not for its neighbours.

The conversation this list is for

There is a question the API specification world has never really answered together: what is commercial open source, and how do the people maintaining the substrate get paid without selling it to whoever shows up with capital?

Several rounds of acquisition in this space have made the failure mode familiar. Something gets built inside a company, becomes load-bearing for other people, and then the company’s priorities change — through acquisition, a strategy shift, or simple attrition — and everyone downstream discovers they were depending on somebody else’s roadmap.

The people on this page are the ones who would have to be part of any real answer. If you are on it, or should be, the conversation is open — and the specific structural question of earmarked sponsorship versus foundation membership is an open issue.

Corrections and additional repositories welcome — the repository list is at the top of the generating script.