Agent API

Displacement data, machine-readable

Every score on displacement.ai is available as static JSON and plain text — no keys, no accounts, no rate limits beyond normal hosting. Built for agents, researchers, and career tools that need task-level AI displacement signals.

Endpoints

Five stable URLs

GET

/api/occupations.json

The full dataset: scores, risk bands, wage context, task signals, and transition pathways for all 620 tracked occupations.

curl https://displacement.ai/api/occupations.json

GET

/api/occupations/{slug}.json

One occupation per endpoint, including task-level automation signals and safer transition paths.

curl https://displacement.ai/api/occupations/data-entry-keyers.json

GET

/llms.txt

Topic map of the site for language-model crawlers: core pages, primary topics, and key URLs.

curl https://displacement.ai/llms.txt

GET

/llms-full.txt

Full plain-text digest: every occupation with scores and summary, every category, and every glossary term in one fetch.

curl https://displacement.ai/llms-full.txt

GET

/ai-workforce-displacement-2026-chart-data.csv

Chart data behind the 2026 workforce displacement brief.

curl https://displacement.ai/ai-workforce-displacement-2026-chart-data.csv

MCP server for agents

The dataset is also available over MCP (Model Context Protocol), so agents like Claude, Cursor, or custom assistants can call displacement.ai as a tool: look up occupation risk, compare roles side by side, list jobs by risk band, and fetch safer transition paths.

Hosted endpoint (no install, Streamable HTTP): https://displacement.ai/api/mcp — add it as a remote MCP server in any compatible client. A local stdio version and full setup notes live at github.com/crosstabs/displacement-ai/tree/main/mcp

How to read the scores

All scores are 0-100 planning signals from a versioned model. Displacement pressure combines task exposure, automation potential, augmentation potential, wage vulnerability, and transition feasibility. Risk bands: very high (70+), high (55-69), moderate (35-54), low (below 35). The model label and review date ship inside every JSON response.

Free to use with attribution to displacement.ai. Scores are planning signals, not forecasts — confirm local wages, hiring demand, credentials, and licensing before making career decisions. See the methodology for the full model description and its limits.