Career data, built for institutionsconsent-first and audit-ready
Book a walkthroughWorks with the agents your people already use
Claude
ChatGPT
Gemini
Cursor
Devin
ManusCohort analytics without raw data
See distributions, shifts, and readiness across a cohort — aggregates only. Raw answers never leave the consent scope.
Consent admin for the whole org
Invite members, scope what agents may read, revoke in one place. Every read is logged and attributable.
API, MCP, and audit by default
Career Context API with org scopes, MCP for agent clients, SSO-ready sign-in, and exportable audit trails.
{ "task": "career",
"member": "org_9f2…",
"dimensions": 22,
"schemaVersion": "1.0" }
Privacy is the product
Built from the ground up so institutions can say yes to career AI.
Read the privacy policy →Consent first
Members grant, narrow, and revoke what agents may read — per scope, any time.
Data minimization
Agents only ever receive derived, consented fields — never raw answers.
Revocable & portable
Delete a share and the agent's access dies with it. Export everything as JSON.
Role-scoped access
Coaches see aggregates; admins see consent state; nobody browses raw answers by default.
Audit trail
Every API read is logged: who, which fields, when, under which consent version.
Retention controls
Org-defined retention windows with scheduled purge, enforced at the storage layer.
What pilot programs ask us
Can our staff see assessment results without exposing raw answers?
Yes — context packets carry dimension scores, versions, and provenance. Raw answers never leave the profile.
How is consent scoped?
Grants carry explicit scopes (profile, assessments, evidence), expire after 7, 30, or 90 days, are revocable — and every read lands in an audit log.
Does it work with our internal agents?
The MCP server and the REST endpoint integrate with standard clients — Claude Code, Cursor, ChatGPT connectors, and custom agents.
What happens to saved results when scoring changes?
Instruments are versioned and saved results keep the version they were scored under — a scoring upgrade is an upgrade, not a re-survey.
Common questions
No by default. Institutions get aggregates and consent state; raw answers stay with the member.
Agents read through the Career Context API or MCP with member-granted scopes; every read is logged.
Scoped data stops being served immediately, and revocation is recorded in the audit log.
Not in v1 — instruments must be taken in-product so raw answers and versioning stay consistent.
Per-seat plans plus pooled AI credits; enterprise agreements add SSO and audit export.
Not yet. Accounts require adult consent today; a minors program is on the roadmap.
For your whole cohort,
roll out the profile layer.
Onboard in bulk
Everyone completes the same seven assessments
Org-level consent policy
Reading scopes set by role
Audit and retention
Full read logs, deletion on schedule