Data handling
What the API keeps, what it does not, and what your key can reach.
| Operation | Retention | What that means |
|---|---|---|
| Parse, Evaluate, Translate, Detect language, Interview summary, Transcribe (synchronous calls) | Stateless | Nothing you send is stored, apart from a short excerpt kept when a model call fails, and no information carries from one request to the next. |
| Register a job | Job description | Only the description you send (with its label and language) is stored, so it can be reused by job_id. Never a candidate, a CV or a result. The job_id is scoped to the account that created it. |
| Prompt (résumé assistant) | Résumé | Keeps the résumé so it can act on it across a session. Browse saved résumés with GET /prompt/resumes. |
Background tasks (/{op}/async) | Task result and uploads | Keeps the task and its result so you can poll for it. Files you upload to a task are stored under your account. |
| Screening interview | Invites, answers, recordings | Keeps candidate invites, answers and recordings so your ATS can retrieve completed recordings and transcripts. |
Your data stays yours
Keys are scoped to your account and see nothing else.
- A task, a saved résumé or a screening is readable only by your account.
- Parsed output preserves the CV's own language and script, so no name or employer is transliterated without you asking for it.
- Every Console view is scoped to your signed-in account. No request accepts an account identifier from the caller.
Testing on production
Production is the only published environment. The synchronous operations keep nothing apart from a short excerpt of a failed model call, so integration testing with them runs safely against production with your own documents. Background tasks, saved résumés, registered jobs and screening interviews do keep data; test those with documents you are happy to have stored.
You remain the data controller. CVs and interview recordings are personal data. Confirm you have a lawful basis for processing candidate material and for sending it to a processor before you integrate.
This page documents how the product operates. It is not a substitute for the Curriculo Terms of Service, Privacy Policy or Data Processing Agreement.
Common questions
Is my CV stored when I call parse or evaluate? No. The synchronous operations are stateless: the document is processed and nothing is kept once the response is sent, apart from a short excerpt of a failed model call.
Which operations keep data? Register a job, the résumé assistant (prompt), background tasks and screening interviews. Each one is listed in the table above with what it keeps.
Can another account read my tasks or jobs? No. Tasks, registered jobs, saved résumés and screenings are scoped to the account that created them. A request for another account's id returns 404.
Where do I get the legal terms? Read the Terms of Service, Privacy Policy and Data Processing Agreement. Questions go to hello@curriculo.me.
Next steps
- Authentication: how keys are issued, scoped and revoked.
- Developer Console: manage keys, usage and billing.
- Background processing and callbacks: what a task keeps and how you poll it.
Updated