Most tools run a single syntax check, call it "verified", and charge you anyway. We run a 7-layer pipeline — syntax, DNS, SMTP, pattern scoring, catch-all detection, deduplication, and proprietary signal enrichment — before a result is ever marked valid. If we can't confirm it, you don't pay.
The pipeline
All 7 layers run sequentially. An address that fails early is dropped immediately — no wasted compute, no inflated credit charges. You only pay when something passes all 7.
Before any network request fires, we validate RFC 5321/5322 compliance — local-part structure, domain format, length limits, and common typo patterns (transposed TLDs, missing dots, double-@ artifacts). Malformed addresses are rejected immediately at zero cost.
We query DNS for MX records to confirm the domain actively accepts email. Domains with missing, misconfigured, or parked MX records fail here. We also detect whether the domain is a known spam trap, disposable address provider, or role-based alias (info@, admin@, noreply@).
We connect to the receiving mail server and simulate delivery — stopping just before the message body is sent. The server's response tells us whether the specific mailbox exists and is actively receiving. A 250 response is a mailbox confirmation; everything else is graded accordingly.
Every domain has an email format pattern (firstname.lastname@, f.lastname@, etc.). We cross-reference the candidate address against known patterns for that domain, then assign a confidence score from 0–100. Low-confidence addresses don't pass — they're surfaced with their score so you can decide.
About 30–40% of B2B domains are configured to accept all incoming addresses — valid or not. SMTP alone can't catch these. We probe with known-invalid addresses to identify catch-all configurations, then apply additional scoring to separate likely-valid from likely-invalid addresses in that domain.
Beyond SMTP, we cross-reference multiple independent external signals — things like whether the address has activity across the broader web, historical reputation data, and engagement patterns. These signals don't change the SMTP result, but they provide a second opinion that catches addresses SMTP would pass but that bounce in practice.
Before a result is written to your account, we check it against your existing Contact DB. Duplicates are marked deduped and returned at no charge. The quality gate also enforces your list's accepted email status thresholds — if you've said VERIFIED only, anything below that threshold is excluded, not charged.
The specific signals and sources behind layers 4–6 are proprietary. We don't publish them because publishing them would allow bad actors to engineer around them. What we do publish: our accuracy numbers, our bounce rates, and our billing model — all independently verifiable against your send outcomes.
Why it matters
Gmail and Outlook use bounce rate as a spam signal. Once your sending domain hits ~5% hard bounce, inbox placement drops. Above 10%, you're looking at blacklisting. Most data tools give you email addresses — not verified email addresses.
Our 7-layer engine was built specifically because SMTP-only verification still misses ~8–12% of invalid addresses (catch-alls, patterns, and reputation). Layering independent signals closes that gap.
| Attribute | LeadTale | Industry avg. |
|---|---|---|
| Bounce rate | <2% | 8–15% |
| Catch-all handling | Layer 5 scoring | SMTP pass-through |
| SMTP verification | Live handshake | Varies (many skip) |
| Failed lookup charge | Free | Often billed |
| Signal layers | 7 | 1–3 |
| Confidence score | Returned per result | Rarely exposed |
Credit model
Running all 7 layers costs real compute. We absorb the cost on failures — because our model only works if you trust the results. If we charge you whether or not we find anything, we have no incentive to be honest about what we can actually verify.
This is not a common billing model in the industry. Most tools charge per lookup regardless of outcome. We charge per verified hit.
See full credit economicsFAQs
Free tier gives you 100 credits every month — no card required. Upload a CSV or run lookups one at a time. You'll see verification status, confidence score, and exactly why each result passed or didn't.