I build with AI.
Ground your agents and apps in real definitions of work, copied from the sources and kept in one place. Open data, open code.
Every way people make a living, written down.
The Org.AI Foundation is a non-profit building an open record of human work: 1,016 occupations and 18,796 tasks, each traced to its source and free for anyone to use.
Every occupation in the record, one mark each. Bigger marks hold more tasks; lit marks match your search.
Make work visible.
Every mark here is one occupation: somebody’s rent, somebody’s craft, somebody’s Tuesday morning.
Written down, work can be seen, checked and argued about. Left unwritten, it changes and nobody notices.
Each dot is still a job. Point at one to see which.
Work is changing faster than anyone is writing it down.
AI is already taking on parts of many jobs. But there is no shared, open record of what those jobs are made of, so nobody can say clearly what is changing, for whom, or how fast.
We are building the record as a public good, free and in the open.
23 families of work.
Sort the record by the kind of work and it falls into 23 families. Production is the largest, with 114 occupations. Legal is the smallest, with 8.
Open one up.
A dental hygienist’s job, as the record holds it, is 16 tasks. Fourteen are core to the job, from “Expose and develop x-ray film” to “Chart conditions of decay and disease for diagnosis and treatment by dentist.”
Every one of the 1,016 jobs opens up like this, into the work it is actually made of.
Every task says where it came from.
Nothing in the record is made up. Each task is quoted from the body that publishes it, and keeps its address, so anyone can check it.
Feel and visually examine gums for sores and signs of disease.
- Source
- O*NET 29.0, U.S. Department of Labor
- Record
- 29-1292.00, Dental Hygienists
- Kind
- Core task, copied verbatim
Next: what a machine can now do.
With the work written down, the next pass goes through all 18,796 tasks, one at a time, and marks which ones a machine can now do as well as a person.
In the open, with the method published, so a worker, a reporter and a regulator can all check the same answer.
Built in the open.
Everything we publish is free to use and free to build on. Start with what is open today, then see the whole map it belongs to.
Open today
Use these now. The code behind every one of them is public.
- schema.org.ai
The shared vocabulary
- standards.org.ai
The standards we keep, and the ones we write
- id.org.ai
Sign-in for people and AI agents
- mdx.org.ai
Documents that people and machines can both read
- primitives.org.ai
The building blocks
- github.com/dot-org-ai
27 public repositories
- /llms.txt
The whole map, for machines
The whole map
144 addresses on .org.ai, each with one job: 142 sorted into seven areas, and two at the root, this site and its directory. Five are open today; the rest are on their way.
Find your way in.
One record, many uses. Start from who you are.
I research, report or write policy.
Every task names its source, and every number we make shows its method. Cite it, check it, print it.
I want to know about my job.
Look up any of 1,016 occupations and see the tasks it is made of, in the words of the people who study it.
I want this to exist.
The record is free for everyone. Partners keep it that way, and pay for the next pass.
Copied from 17 sources. Never edited.
We do not write our own version of the facts. We quote the bodies that publish them, and keep the address of every one. Here is one task, traced all the way back.
- The task
Feel and visually examine gums for sores and signs of disease.
- StatementTask 417, a core task, rated by 54 people who do the jobO*NET survey, August 2021
- OccupationDental Hygienists, 29-1292.00One of 96 occupations in Healthcare Practitioners
- ClassificationStandard Occupational Classification 2018, 29-1292The U.S. government's codes for every occupation
- Published byO*NET 29.0, U.S. Department of LaborEmployment and Training Administration
- LicenseCC BY 4.0Free to copy and reuse, with credit
- Changed by usNothing.Quoted as published, with its address kept
All 17, by name.
Grouped by what they describe. Each one is copied exactly as its publisher wrote it.
Work
- O*NETOccupations, and the tasks inside themU.S. Department of Labor
- SOCThe standard codes for occupationsU.S. Bureau of Labor Statistics
- APQCHow business processes are named and groupedAmerican Productivity & Quality Center
Industry and trade
- NAICSThe standard codes for industriesU.S. Census Bureau
- GS1Barcodes and the classification of productsGS1, the global barcode standards body
- EPCISHow goods are tracked through supply chainsA GS1 standard
- X12Electronic business documents, like orders and invoicesAccredited Standards Committee X12
- UNInternational codes and classificationsUnited Nations
Health
- HL7How health records are exchangedHL7 International
The web and its standards
- ISOInternational standards, from country codes to currenciesInternational Organization for Standardization
- IEEEEngineering and computing standardsInstitute of Electrical and Electronics Engineers
- IANAThe internet’s registries: domains, media types, time zonesInternet Assigned Numbers Authority
- W3CThe standards the web is built onWorld Wide Web Consortium
- MarkdownThe plain-text way people write for the webJohn Gruber
- MCPHow AI agents connect to tools and dataModel Context Protocol
Knowledge
- WikipediaThe free encyclopediaWikimedia Foundation
Measurement
- GDPvalHow well AI does real, paid work tasksOpenAI
And our own receipts.
Where we make a number ourselves, we publish how: the method, the exact file we loaded, and where it came from. It is all at standards.org.ai.
Keep the record open.
The record is free for every worker, student, reporter and builder who needs it. It stays that way because people and institutions choose to support it.
Donate
Your gift keeps the record free, and pays for the next pass: marking, task by task, what a machine can now do.
Partner with us
Fund a part of the record, contribute data, build on it, or study it. Tell us who you are and what you have in mind.