Validation policies and lists
The result verdict is the same for every customer. Two tools, set in the dashboard, adapt it to your rules: the validation policy, which produces finalDecision, and the allowlist and blocklist, which settle the matter for specific addresses or domains.
Validation policies
A policy (Validation rules in the dashboard) lists what you reject. It applies to every call and feeds two fields of the response:
"finalDecision": "rejected",
"policy": {
"decision": "rejected",
"reasons": ["reject_disposable", "reject_role_account"],
"rejectedByClientConfig": true
}
| Rule | Code in policy.reasons | Rejects when |
|---|---|---|
| Reject disposable addresses | reject_disposable | features.disposable flags a disposable address |
| Reject malicious domains | reject_malicious_domain | features.maliciousDomains flags the domain |
| Reject trap domains | reject_typosquat | features.typosquat flags the domain |
| Reject generic accounts | reject_role_account | contact@, info@, support@… (noreply@ included) |
| Reject noreply addresses | reject_no_reply | noreply@, mailer-daemon@… only: contact@ is still accepted |
| Reject accept-all domains | reject_catch_all | The domain accepts any address |
| Reject consumer providers | reject_free_provider | Gmail, Outlook, Orange… (B2B forms) |
| Reject inconclusive results | reject_unknown | result is unknown |
| Reject offensive terms | reject_on_profanity | features.profanity flags a term |
| Reject random local parts | reject_on_random_detection | features.randomDetection flags it |
| Minimum score | reject_low_quality_score | score below the chosen minimum |
| Minimum domain health | reject_low_domain_health | Domain health below the chosen level (excellent → critical) |
| Reject VPN and Tor | reject_vpn_or_tor | The call comes from an anonymising network |
| Reject hosting providers | reject_datacenter_ip | The call comes from a hosting provider |
finalDecision does not replace resultThe policy evaluates only its own rules. An invalid address that no rule targets comes out with finalDecision: "accepted". Without a policy, finalDecision is always accepted. So decide on both: reject if result is invalid or if finalDecision is rejected.
The reject_vpn_or_tor and reject_datacenter_ip rules look at the IP that calls the API. On a server key, that is your own server, hosted in a data centre: reject_datacenter_ip would then reject every address. Keep these rules for widget policies.
Which policy applies
- The policy linked to the API key, if the key has one.
- Otherwise, the policy marked as default in the organisation.
- Otherwise, none: everything is
accepted.
Create one policy per use (B2C form, B2B form, newsletter widget…) and link it to the matching key.
Allowlist and blocklist
Exception lists in the dashboard. Each entry targets a specific address or a whole domain.
| List | Effect on verification |
|---|---|
| Allowlist | SMTP does not run (no credit). The score is at least 95. The verdict follows syntax and DNS. |
| Blocklist | SMTP does not run (no credit). The verdict is invalid, with features.listCheck.details.isBlacklisted: true. |
The same value cannot be on both lists: adding it is refused with the code LIST_CONFLICT and the dashboard offers to move the entry.
The lists also apply to batches and to the widget. A key can ignore them by disabling its whitelist or blacklist checks.