AI & agent access policy
How AI systems and agents may use EconomicsAPI's 50 countries and 23 endpoints of data.
Citation is welcome
Any AI system — a search assistant, a research agent, a chatbot answering a factual question — may cite figures from this site freely, with attribution back to EconomicsAPI (and, where a page names it, the original national or international source). This applies to individual facts and short excerpts used to answer a specific question.
Use the API or MCP, not scraping
Every number on this site is also served as structured data: 23 REST endpoints, one MCP tool per endpoint. If an agent needs to read data programmatically — one value, a full series, or many countries at once — it should call the API or connect over MCP rather than parse the HTML pages. Structured access is faster for the agent, lighter on this site, and less likely to misread a table. See llms.txt for the full catalogue of pages and endpoints, machine-readable.
Bulk extraction and training need consent
Systematically crawling or reproducing this site's content at scale — mirroring the full dataset, or using it as training data beyond incidental inclusion in a general web crawl — requires our consent first. Contact us to talk about a license or a data partnership.
What crawlers see
Our robots.txt opens the site to named AI search and training crawlers and states this policy as a machine-readable Content-Signal header. Junk crawlers that ignore robots directives are blocked outright.