Bring the workforce to your AI.
Every chart on workforce.ai is a call you can make yourself. Connect us to Claude or ChatGPT and your own AI asks the questions, or call the API from code. Same data, any paid plan. Your AI's calls count against the plan's daily calls, never against your questions.
| Plan | Calls a day | Names |
|---|---|---|
| Free | None | None |
| Basic, $5 | 500 | 250 a month |
| Pro, $20 | 2,000 | 1,000 a month |
| Max, $99 | 10,000 | 5,000 a month |
Names come out of the same monthly names as on the site. Questions to our own AI count against your plan's questions; your AI asking us through the connector does not.
In Claude or ChatGPT
Add workforce.ai as a connector and sign in with your workforce.ai account. Then just ask: "Where did Stripe's data scientists go this year?" or "What does Austin's workforce over-index on?" Your AI picks the tools, reads the numbers and writes the answer.
Connector address https://mcp.workforce.aiClaude, ChatGPT, Cursor, VS Code and Claude Code each get their own steps there, with your key filled in when you're signed in. Apps that take a header can send Authorization: Bearer wf_... instead of signing in.
What your AI can reach
find_company | Companies by name or ticker, biggest first, group and entity |
company_genome | One section of a company: vitals, where joiners came from, where leavers went, departments, levels, schools, tenure, hiring classes, promotion pace, leaders, locations, relatives. Filter by department, level, title, country, school |
compare_companies | Vitals for two to ten companies side by side |
people | How many people match a slice and the newest 10, free |
people_list | The full list, from your plan's monthly names; says the cost first |
job_changes | Every move between employers by month, for a company or the whole panel |
recent_job_changes | Changes spotted day by day over the last 30 days |
college, city | Where a college's graduates work; what a metro's workforce over-indexes on |
Every answer says its window and what its shares are of, so your AI can't pass off a slice as the whole.
The API
Get your key from your account. Send it as a bearer token.
A company, by section
curl -H "Authorization: Bearer wf_..." \
"https://workforce.ai/api/genome?co=Salesforce§ion=summary"Sections: summary (monthly headcount, joined, left, twelve-month facts), levels, dna (departments), tenure, locations, schools, ancestry, arrivals (where hires came from), departures (where leavers went). Filter with fn (department) and lv (level), and set the window with months.
Up to ten companies at once
curl -H "Authorization: Bearer wf_..." \
"https://workforce.ai/api/group?cos=Salesforce,Workday,ServiceNow§ion=summary"The people behind a number
curl -X POST -H "Authorization: Bearer wf_..." -H "Content-Type: application/json" \
-d '{"action":"sample","spec":{"co":"Stripe","kind":"left","fn":"Engineering","from":"2025-07","to":"2026-06"}}' \
https://workforce.ai/api/peoplesample returns the count and the newest 10. quote says how many names the full list takes from your monthly names; download takes them and returns the file.
Rules of the data
A panel, not a census: counts are people in our sample. The newest months fill in as people update profiles. Breakdowns return the 25 largest groups. Read how the data works before you build on it.