# The Org.AI Foundation · Standards > We copy other people's records and never correct them. The few we wrote ourselves sit > on a second shelf, and say how unfinished they are. standards.org.ai is the register of the Org.AI Foundation. It holds two shelves. The first is other standards bodies, copied into this estate and never edited by us. The second is the small number of standards this Foundation writes itself. Nothing here is generated at request time and nothing here updates itself. ## Status Everything below was measured from the standards.org.ai repository on 2026-07-28, at git HEAD 08ee939. Every figure is followed by how it was counted. Nothing is estimated, rounded, or carried over from an earlier document. 18 top-level folders on disk. 43 collections inside them hold at least one row. 466 original source files are kept in the repository. 223 generated entity files and 92 relationship files. The arithmetic that proves the census, by section, from .data/*.tsv with headers excluded: Work 172,048 Made and sold 202,790 Health 117,184 Money and paperwork 5,183 The lists everyone quotes 124,375 Government registers 5,925 --------------------------------- first shelf 627,505 ours, not any body's 514 --------------------------------- total rows 628,019 628,019 is the raw row count. It is not 628,019 distinct things: the .Scored, .Extended and .Metadata views restate the same referent with extra columns, and one pair of files is byte-identical. Counting unique (namespace, type, id) triples across the 223 files, with id falling back to code where absent, the distinct figure is 478,507. Relationship rows: 861,416. The 514 rows that are ours are Graph.Domains, Graph.Routing, Graph.Types, Superset.Skills and Superset.SkillCategories. They belong to the second shelf or to nothing, and must never appear in a count of what was copied from outside. Two things are open and are marked as such wherever they appear: - How much of each collection we hold has not been measured against any publisher's own total. No share, ratio or percentage is printed on any surface until it is. - Nothing on the second shelf has an external binder. None of the six is at 1.0 and none will be until at least two people outside this Foundation have built something that depends on it. ## The machine face There is no API here. There is no download endpoint, no search endpoint, no query interface and no feed. The data lives in a git repository and this file is the only machine face standards.org.ai currently serves. Do not treat any name in this file as a resolvable endpoint. As of 2026-07-28 the hosts that serve a real page a person can read are: axp.org.ai, mdx.org.ai/docs, mdxld.org, product.md, schema.org.ai, todo.md.org.ai, agents.md.org.ai, foundation.org.ai. onet.org.ai and apqc.org.ai serve holding pages. wikipedia.org.ai returns JSON and no human page. soc.org.ai serves another site's page. naics.org.ai, un.org.ai, x12.org.ai, hl7.org.ai, iso.org.ai, ieee.org.ai, iana.org.ai, w3.org.ai, epcis.org.ai, markdown.org.ai, mcp.org.ai and gdpval.org.ai do not respond. ## The first shelf: what we copied Six sections. Each names a body, what it governs, what it feeds in this estate, the version and date we loaded, and how much of it we hold. Rows per collection are from .data/*.tsv grouped by filename prefix, headers excluded. ### Work (7 collections, 172,048 rows) O*NET 133,288 database 30.0, folder db_30_0_text. Public domain. Occupations 1,016; tasks 18,797. APQC PCF 30,987 7.4 Cross-Industry plus the industry frameworks at 7.2.1 and 7.2.2. Membership or purchase; the file grants a royalty-free licence with attribution. Advance CTE 3,978 Framework Crosswalk, September 2025. Free with attribution. BLS 3,447 OES May 2024; projections on 2022 NAICS, 2018 SOC. Public domain. One large series file undecoded. CEDS 158 No version stated. A sample of its elements. ISCED 137 ISCED 2011 levels, ISCED-F 2013 fields. UNESCO. CASE 53 No version stated. Frameworks and types only. ### Made and sold (7 collections, 202,790 rows) UNSPSC 183,288 No version stated. Registration required to download. No page templates: held as columns only. GS1 8,804 Product classification as of May 2025 with the November 2024 network release; events at 2.0.1; web vocabulary at 1.17. GS1 licence, free for members. NAPCS 5,362 NAPCS 2022. Public domain. Schema.org 3,040 No version stated. CC BY-SA 3.0. NAICS 2,142 NAICS 2022. Public domain. ETIM 100 No version stated. 80 example classes of the several thousand it publishes. eCl@ss 54 No version stated. 49 segments, and none of the 40,000-plus classes its own documentation names. ### Health (9 collections, 117,184 rows) ICD-10-CM 98,208 FY 2026, 1 Oct 2025 to 30 Sep 2026. Public domain. NDC 10,000 Downloaded 5 Dec 2025. A 10,000-product sample of a directory with 135,586-plus in it. Public domain. HCPCS Level II 7,907 January 2026. Public domain, CMS. NPI provider taxonomy 883 Version 25.1, released 1 Jul 2025. NUCC. HL7 FHIR 127 R5, v5.0.0. CC0. Resources and datatypes. SNOMED CT 20 Top-level hierarchy only. Full set needs a SNOMED International licence we do not hold. RxNorm 16 Monthly release, October 2025. Term types only. LOINC 13 2.81, released 12 Aug 2025. Axes and categories only. LOINC licence agreement. CPT 10 Category information only. Codes are the AMA's and need a licence we do not hold. ### Money and paperwork (9 collections, 5,183 rows) X12 1,298 No version stated. 318 transaction sets, 845 segments, 135 elements. LEI 1,123 No version stated. Registration authorities. GLEIF. MCC 1,010 ISO 18245:2023. ISO TC68 and the card networks. ISO 20022 657 No version stated. Message catalogue. ISO TC68; registration authority SWIFT. EANCOM 278 No version stated. GS1. BIC / SWIFT 254 ISO 9362:2014. Structure and country codes only; not the register of banks. Peppol 236 No version stated. OpenPeppol. UN/EDIFACT 207 Directory D.20B. 195 of the 317 message types it publishes, plus 12 categories. UN terms of use. ISIN 120 No version stated. Numbering agencies. ANNA. ### The lists everyone quotes (5 collections, 124,375 rows) UN/LOCODE 120,807 Notes cite the loc241 release. UN terms of use. W3C 2,077 22 specifications. Per-spec versions not stated. W3C Software and Document Licence; MDN licence cited for some tables. Four of the 22 folders are not W3C's: DOM and HTML are WHATWG's, Dublin Core is DCMI's, FOAF is the FOAF project's. ISO 3166 / 4217 / 639 915 No edition stated. Some ISO standards are free and others must be bought. IANA time zone database 547 From the vvo/tzdb mirror, downloaded 7 Dec 2025. Public domain. UN M49 29 No version stated. Regions only; no country rows. ### Government registers (6 collections, 5,925 rows) Census Bureau 4,148 FIPS codes with the July 2023 metro-area delineation, downloaded 7 Dec 2025. Public domain. SEC 1,095 Four-digit SIC list, downloaded 7 Dec 2025. Public domain. GSA 228 PSC Manual April 2025 with the DLA supply classes. Public domain. Classes only. USITC 219 Downloaded 7 Dec 2025. Public domain. Chapters only, nothing beneath them. USPTO 203 Nice 12th edition 2025, with Locarno, CPC and USPC, downloaded 7 Dec 2025. Public domain. Nice's 45 and Locarno's 32 classes are whole; CPC and USPC are a handful of rows. SBA 32 13 CFR 121.201. Public domain. Business and contract types; not the thresholds by industry. ### Collections held only in part Twelve collections hold a fraction of the standard they name, and each says so on its own row on the page: CPT, LOINC, RxNorm, SNOMED CT, SBA, CASE, eCl@ss, ETIM, CEDS, USPTO, GSA, USITC. NDC, UN M49, UN/EDIFACT, BLS and BIC/SWIFT are partial in the ways named above. ### The publisher count, and why the page does not print it Counting distinct publishing organisations named by this repository's own READMEs, and merging collections that share a publisher (GS1 publishes both GS1 and EANCOM; ISO publishes both ISO and ISO 20022; the U.S. Census Bureau publishes NAICS, NAPCS and the Census geography; CMS publishes HCPCS and NDC), the figure is 37. It rests on attributions rather than on a file count, so the visible page prints 18, 43 and 466, which are mechanical, and leaves 37 here. ## Reserved but empty This Foundation reserved a name in its estate for seventeen outside bodies. Eleven of those names have data on disk today: onet, apqc, naics, gs1, un, x12, hl7, iso, iana, w3, epcis. Six have nothing loaded: soc, ieee, wikipedia, markdown, mcp, gdpval. SOC codes appear inside the BLS and Advance CTE files but there is no SOC collection. Seven folders that hold real data answer to no reserved name at all: Advance CTE, BLS, Ecommerce, Education, Finance, NAPCS and the U.S. agency folder. Between them they hold 25,418 rows. The honest one-line version: 18 folders on disk; 43 collections with data; 6 of the 17 reserved names still empty. ## The second shelf: what we wrote Six. None is at 1.0. None has an external binder. AXP — the Agent eXperience Protocol. What an API must publish so an agent can transact with it on first contact with no human reading documentation. Draft 0.3.0, dated 2026-07-21. Publisher apis.ax, verifier api.qa. $id https://apis.ax/axp. Home: https://axp.org.ai MDXLD — MDX carrying $id, $type and $context in its frontmatter. A superset of the open MDX format whose frontmatter is a superset of YAML-LD. $context is permanently https://schema.org.ai and carries no version. No version number of its own. Home: https://mdx.org.ai/docs and https://mdxld.org PRODUCT.md — one Markdown file describing who a product is for, how it speaks and what it costs. Draft; the owner's ruling records 0.9 and the file itself carries no version string. Its own page states that it is a specification we write, not a standard we host. Home: https://product.md TODO.md — a task list a machine can also read, in a plain Markdown file. No version number. In use in this Foundation's own repositories. Home: https://todo.md.org.ai schema.org.ai — our superset of Schema.org, extended for work machines do. No version number. CC BY-SA 4.0; the vocabulary it extends is CC BY-SA 3.0. Its README states 871 types and its things directory holds 905 pages; the two have not been reconciled, so no total is printed anywhere. Home: https://schema.org.ai AGENTS.md frontmatter profile — the frontmatter an AGENTS.md file carries so one source compiles to AGENTS.md, CLAUDE.md and .cursorrules. No version number. Home: https://agents.md.org.ai ## The counters Copied from others: 18 folders, 43 collections with anything in them, 466 original files kept. Git-tracked top-level folders, excluding the dot-directories, node_modules and site/; collections holding at least one row in .data; git ls-files .source. Written down: 478,507 entries and 861,416 links between them. Unique (namespace, type, id) triples across the 223 entity files, id falling back to code; rows carrying neither an id nor a code are not entities and are not counted, and counting them would print 478,709 instead. Relationship rows across the 92 relationship files. The raw entity row count is 628,019 and the difference is the same referent restated in more than one view. How much of each: not measured, and therefore not printed. No share, ratio or percentage for any collection appears on any surface of this site. ## How a page says where it came from Every entry page in this estate carries one line of five plain phrases: Seeded from, Curated here, Built as, Minted at, Operated by. On a page on the first shelf the second phrase reads "Curated at" and names the curating home, because the copy is not the curation, and the last three are blank and stay blank: a record we copied was not built, minted or operated by anybody here. Where a step did not happen the line shows a visible blank rather than a plausible name. A second line carries the loading receipt: the source file, the licence, the generated file the row landed in, and the publisher's own code kept unchanged. ## Licence and use The transformation in this repository is CC BY-SA 4.0, copyright 2025 Standards.org.ai. That covers the transformation and not the sources. Each source keeps its own licence and each is named above; several require a licence or a purchase from their publisher before the full set may be used, and those are marked. We have not contacted any of the bodies named in this file. Nothing here is published with their endorsement, their participation, or their knowledge, and nothing here should be read as a claim of any of the three. Nothing on standards.org.ai is generated at request time. This page and this file are static, self-contained, and make zero external requests.