Methodology

Exactly how we verify an address, score it, and what our performance figures mean.

How a verification works

Each verification runs a fixed sequence of checks and stops early when a result is already certain:

  1. Syntax — the address is parsed against a practical RFC 5322 grammar (valid local part, one “@”, a real domain with a TLD). Malformed addresses are rejected here with no network calls.
  2. Domain & MX — we resolve the domain’s MX records (falling back to an A record). No mail exchangers means mail can’t be delivered, so the address is undeliverable.
  3. Live SMTP mailbox probe — we open an SMTP connection to the highest-priority MX and issue MAIL FROM / RCPT TO to ask the server whether the mailbox exists — without sending a message. We read the real response code (250 accept, 550 no-such-user, 4xx greylist, etc.).
  4. Catch-all detection — we also probe a random address at the domain. If it’s accepted too, the domain is “accept-all,” so a positive result for the target no longer proves the mailbox exists.
  5. Quality signals — disposable/throwaway domain, role account (info@, support@…), free consumer provider, gibberish local part, and a “did you mean” typo suggestion.

How the score is calculated

The individual results are fused into a single 0–100 confidence score and one of four statuses: deliverable, undeliverable, risky (accepts but low-quality — catch-all, role or disposable), or unknown (the server couldn’t be reached: greylisting, timeout or a blocked port). The score reflects our confidence in that status, not a guarantee — you choose the threshold that matches your risk appetite.

What we cannot determine

  • Catch-all mailboxes. On accept-all domains, SMTP acceptance is not proof — use engagement or double opt-in.
  • Greylisting / blocks. Some servers defer or block unknown probes; we report these as unknown rather than guessing.
  • Outbound reputation. A domain scan checks inbound routing and DNS — it doesn’t measure your sending IP’s reputation. See the Email Health Check notes and the (upcoming) Outbound Deliverability Test.

Performance figures

Our homepage cites a 99.9% uptime target and a typical verification response under 500 ms. These are operational targets and typical observed values, not a contractual guarantee. Live status is on the status page; formal commitments are described under service levels. Actual latency varies with the destination mail server’s responsiveness and greylisting.

Diagnostic tools

The DNS and deliverability tools (MX, SPF, DKIM, DMARC, MTA-STS, TLS-RPT, BIMI, blacklist, SMTP, headers) query authoritative/public DNS and the target servers directly. Where a data source blocks automated queries (e.g. some blocklists), we report “unavailable” rather than assume a clean result. The DNS/SMTP diagnostic checks are also available as diagnostics API endpoints (the client-side generators and validators run in your browser and have no server endpoint).

Last reviewed: 2026-07-19.