VerifAi documentation

VerifAi verifies whether an email address can receive mail and cleans your contact data in a single pass. These pages explain what our results mean, how we prove them, exactly what you pay for, and how to call the REST API — everything a technical buyer needs before signing up.

The REST API

Two endpoints, one bearer key, one JSON object per call.

QuickstartCreate a key and verify your first address in two steps, with a copy-paste example.AuthenticationAPI keys, rate limits per key, and safe retries with idempotency keys.API reference/v1/validate and /v1/credits — full schema, real responses, curl / Python / JavaScript.

The reference is also published as a machine-readable OpenAPI document, maintained by hand alongside the endpoints.

Core ideas

Every check produces three things: a verdict (can you mail this address?), a machine-readable reason (why), and per-field provenance (how we know). You choose how deep to probe with a depth tier, and you only pay for answers.

Verdicts & reasonsThe four verdicts, every reason code, and what to do about each.ProvenanceThe six origin values — and why checked is not the same as certain.Depth tiersSyntax, MX-only and full — what each proves and what each costs.How billing worksThe seven billing rules, stated plainly. Unknowns are free.Catch-all resultsWhy a successful probe against a catch-all domain proves nothing.EvidenceThe chronological on-the-wire record behind every full check.ErrorsThe error catalogue, with stable anchors.

These docs describe behaviour that is stable regardless of how you call VerifAi — through the app or the REST API. Where a page states a billing rule, that rule is what the platform actually does; if you ever see a discrepancy, the running system is the source of truth.

← Back to home