Terms

What is provided

A hosted service for ingesting content and retrieving it with citations, within the limits configured for each organization.

Organizations are provisioned rather than self-served, and the limits — the monthly embedding budget in particular — are part of the arrangement rather than a surprise in a bill.

What is expected

That content ingested is content the organization is entitled to ingest, and that a live dataset points at a site whose owner permits crawling. The crawler honours robots.txt by default; switching that off is the organization's decision and responsibility.

That API keys are kept secret. A key is shown once and acts in one organization; anything done with it is attributed to that organization.

Availability and change

This is pre-release software under active development. Interfaces are versioned from the first line of code, and a breaking change means a new version rather than a quiet replacement.