Privacy
How notanATS handles data, in plain terms.
This page describes the product's actual data practices, as built. It is a plain-language description, not a legal policy — the practices below are the mechanisms the system runs on.
What the product holds
For hiring teams: account details, positions, job descriptions, and the calibration models managers create. For applicants: the application, the resume, the screening conversation, the scores it produced, and — only when the organisation enables it and the applicant chooses to answer — voluntary demographic information used for fairness monitoring.
Masking at the AI boundary
Before screening content is processed by AI, the applicant's name and email address are masked out of the record being screened and scored. The screening works from the applicant's answers. AI processing is performed with Anthropic's Claude models; product data is hosted on Supabase infrastructure; email is delivered via Resend; payments, where they exist, run through Stripe.
No tracking of applicants
Analytics and marketing tracking run only on this marketing website, never on applicant pages or inside the product. An applicant's screening is not instrumented, and no advertising signals are ever collected from anyone. Consent on the marketing site defaults to denied; advertising categories stay denied even after a visitor accepts analytics.
How long data is kept
Retention is enforced by a scheduled purge, not by intention. On free plans, screening conversations are deleted 30 days after they end. On paid plans, applications are anonymised after 12 months: identifiers and resume files are removed, and aggregate scores are kept.
Erasure
An applicant can ask for their data to be erased. Applications still live in an active hiring pipeline are excluded until the process ends; everything else is anonymised in place — identifiers removed, scores preserved in aggregate. There is a route for erasure requests even when the applicant cannot reach the hiring company directly.
Who can reach what
Managers see their own organisation's data. Applicants reach exactly one application, through a private link that expires, is checked against their inbox on first use, and binds to their device. API keys are scoped to one organisation and stored only as hashes. The private scoring rubric is stripped from everything that leaves the system — exports, ATS writebacks, and API responses alike.
Questions
Practices you can ask us to walk through.
Any mechanism on this page can be explained in detail — what is masked, what is kept, what is deleted, and when. That conversation starts with access.
Request access