Field API
For your organisation’s AI agents.
Field API lets your own agent ask what a person asks here: frame a materials problem and get a provisional direction, the methods that fit, constraints and the questions that could change the choice; or name the method families likely to serve a stated need. It returns directions, questions and limits — not a catalogue of laboratories, quotations, bookings or provider contact.
Pilot: by invitation for now. Access is approved by a person, one organisation at a time, and each credential is tied to a named user in a workspace.
Pilot access is arranged directly with approved organisations. It has no public price and is not sold through self-service; our paid offer is case support.
Limits in the pilot
- One task-shaped scientific question at a time, and at most three provisional method directions in reply.
- No listing, wildcard, paging or bulk export of the underlying data, and no way to read the capability graph directly.
- No provider contact, private rate, current availability, quotation or routing action of any kind.
- Every credential is scoped, expiring, revocable and attached to a named user in a workspace; usage and model spend are limited server-side.
A credentialed, task-shaped interface: never anonymous, never a bulk catalogue API.