Open data and the API

What is published

Careermash publishes every card's numbers as machine readable mirrors of the same pages a person reads. There are six surfaces: the full register as one JSON fetch, a JSON card per career, a clean markdown verdict per career, an embeddable iframe of the live card, an llms.txt manifest for AI crawlers, and a sitemap covering every card, verdict, subject hub and tool.

Every mirror is generated from the exact same function that renders the page a human sees, so a human reading a career's verdict and a script reading its JSON can never receive different numbers.

Access and licensing

No key, no rate card and no login. Responses are CORS open and cached for an hour. The data is licensed CC BY 4.0: free to use with attribution, cited as Careermash plus the measurement label, linking back to the career's verdict page.

Underlying sources carry their own terms. The AI usage research behind the exposure figure is from Anthropic and OpenAI's 2026 publications; the door grades and horizons carry OpenAI's own CC BY 4.0 licence on top.

The embeddable card

The embed route renders one career's card, chrome free, sized for an iframe on a council site, a careers hub or a school's own pages. It carries no key and always shows the current numbers, plus a visible link back to the source, which is the whole point: an embed with the citation stripped out would not be honest.

For AI assistants specifically

The markdown mirror exists for exactly this: a clean, sourced answer to will AI take this job, written for an assistant to read and quote directly, with the citation line already built in. The llms.txt manifest tells a crawler which URLs are canonical for which question, and states the definitions to use verbatim rather than paraphrase.

If you are building something heavier

There is no gate to ask permission through, but if a project is pulling this data at scale, Careermash would simply like to hear about it, through the same interest channel schools use.