Embed validated assessments in your product.

It exposes the JobCannon assessment engine as a programmable service your product can call. Your users take a validated assessment — RIASEC, Big Five, EQ, Skills Audit, or any of the 138+ instruments — inside your own app, and you get a scored result back. The full loop is three REST calls: POST /api/v1/assessments starts a session and returns the question set; GET /api/v1/assessments/{session_id}/questions re-fetches those questions if you need them again; POST /api/v1/assessments/{session_id}/complete submits the answers, scores them server-side, and returns the result. The same scoring code path the consumer product uses (lib/scoring/compute.ts) runs behind the API, so an API result is identical to one taken on jobcannon.io.

Every request carries an X-API-Key header. Your key looks like jck_live_… and is issued per organisation with scopes — assessments:send lets you start and complete sessions, assessments:read lets you re-fetch questions. Keys are SHA-256 hashed at rest and never stored in plaintext. There is no OAuth dance and no client-secret exchange: one header, scoped, revocable. A session is a short-lived signed token returned by the start call, so there is no session table to manage on your side and no polling.

POST /api/v1/assessments — body { test, locale?, external_user_id?, pool? }; returns { session_id, test, locale, question_pool, total_questions, questions[] }. GET /api/v1/assessments/{session_id}/questions — re-serves the question payload for an in-progress session (scope assessments:read). POST /api/v1/assessments/{session_id}/complete — body { answers: number[], external_user_id? } where answers are option indices in the order the questions were served; returns { result_id, test, top_result, scores, question_pool }. Answers are validated for length and for degenerate/untouched vectors before scoring. Pooled instruments (e.g. the IQ battery) thread the served question pool through to the right scoring key.

Each completed assessment is persisted to your organisation — the result row is written with your api_org_id, so it is attributed to you, not to an anonymous browser session or a JobCannon consumer account. The complete call returns the result_id, the top_result (the headline archetype/type/code for that instrument), and the full per-dimension scores. You store the result_id and render the outcome inside your product however you like. The current v1 returns the scored result; the richer AI Career Guide narrative (course recommendations, 2,521-career mapping) is a separate consumer surface and is scoped per partner, not a default v1 response.

Requests are rate-limited per API key (the key identity, not the caller IP), so one noisy integration never starves another. Over the limit you get an HTTP 429 with a resetAt timestamp in the body — no silent throttling, no dropped requests. Every start and complete call is metered so your monthly assessment count is accurate for billing.

Teams that build career-assessment in-house usually underestimate two costs. (1) Validated item banks and scoring — the RIASEC, Big Five and EQ frameworks are public, but validated item pools and directional scoring algorithms are not; building and validating your own takes years and tens of thousands of administrations. (2) Maintenance — instruments drift and item banks need refresh. The API is the shortcut: you call three endpoints, we keep the psychometrics current. It is live on the Scale tier ($199/mo), which includes a monthly assessment allotment, then $2 per assessment over the cap. White-label partners who resell our assessments can move to a revenue-share instead.