Skip to main content

Core concepts

A short glossary of the terms used throughout the docs and API.

Methodology​

A versioned assessment definition owned by the Catalog service: the competencies it measures, the ECOs (observable behaviours), and the questions Claire asks. Methodologies are locale-aware — the same methodology UUID serves every language.

Evaluation​

A single assessment run, owned by the Assessment service (user_evaluations). An evaluation has exactly one owner — a B2C individual, a tenant member, or a tenant end-user (widget visitor) — and moves through in_progress → scoring → completed. Locale is fixed at creation.

Claire​

The AI assessment agent (AI Service). Claire conducts the conversation over the text channel (SSE) or the voice channel (WebRTC), and owns the transcript across its lifecycle.

Tenant​

A B2B organisation. Tenants have members (with admin | manager | member roles), branding, adopted methodologies, and may publish widgets.

Widget​

An embeddable assessment surface a tenant places on its own website. The widget authenticates browser visitors with a short-lived visitor JWT minted by the tenant's backend.

Visitor JWT vs. client-credentials token​

Two bearer types reach the public API:

  • Visitor JWT — Ed25519, ~15-minute TTL, minted per visitor session. Used by the widget in the browser. The visitor never holds a long-lived secret.
  • Client-credentials token — OAuth2, minted by an integrator backend from a Wiselook-issued client. Used for server-to-server calls.

See Authentication for the full flow.