Docs

Console: Cookies

Run a scan, cascade classification, review queue and promotion.

In CookiesRun scan, give a start URL and pick the environment. A hosted crawler visits the site, accepts the cookie banner, and reports every cookie and web-storage key it observes — they land in the inventory, already classified. Import scan accepts a JSON array { name, domain, party, expirySeconds? } if you run your own scanner instead.

Classification cascade

  1. Reference DB — a curated dataset of well-known cookies, plus classifications promoted from real reviews. Exact match → confidence 1.0.
  2. Heuristics — name patterns and expiry/party rules.
  3. AI — a model classifies the rest. Anything it is not confident about goes to the review queue.

Review queue & promotion

Classifications below the threshold go to the queue. A HUMAN override is authoritative for that site. When ≥2 independent tenants confirm the same classification, it can be promoted to the shared CookieDefinition — it then resolves at tier 1 for everyone.