AutomationSelf-Hosted

LeadGenEngine

A Zero-Cost-Per-Lookup B2B Email-Discovery Pipeline

A self-hosted pipeline that turns a company domain into SMTP-verified emails for titled employees at effectively zero marginal cost replacing a five-figure annual enrichment-API bill.

PythonFastAPIPlaywrightCamoufoxScrapyhttpxdnspythonMongoDBMotor

Problem Statement

A sales team was bleeding budget on per-record enrichment APIs that metered every lookup, yet still couldn’t reliably reach the SEO decision-makers that mattered. Naive email guessing tops out near a 65% hit rate, and every bounce erodes sender reputation.

Per-lookup API pricing turned prospecting into a tax on growth at 1k–10k lookups/month.
Target SEO persona was under-represented in vendor contributory databases.
Blind permutation hits ~65% accuracy; misses damage domain deliverability.

Headline Outcomes

≈ $0self-hosted

Marginal cost per lookup

Metered (paid API)

75–85%+10–20 pts

Deliverable find-rate

~65%

< 5%

Bounce on high-confidence sends

Uncontrolled

The Solution

A staged, self-hosted pipeline that mirrors a diligent researcher: start from public signals, corroborate independently, and only trust an address once the mail server confirms it with adaptive routing instead of one rigid recipe.

A DNS MX-record gate fingerprints each domain before guessing, rerouting no-mail domains to personal-account discovery.

Discovery beats permutation: harvests real public emails to infer the company’s single naming convention.

A fused 0–100 confidence verdict (pattern + provider + signals) classifies every candidate as accept/review/reject.

On-prem SMTP verification closes the loop output is verified-only, and every confirmation locks the domain pattern.

System Architecture

How the data flows

Company Domain

Input

LinkedIn Page

Titled employees

Email Permutations

Ranked by inferred format

SMTP Verify

On-prem, server-side

Verified Output

Confirmed mailboxes only

Result 01

Replaced a recurring five-figure enrichment-API line item with an owned open-source asset.

Result 02

Closed the SEO-persona coverage gap by sourcing directly from LinkedIn and live search.

Result 03

Self-improving: each verified address locks a domain’s pattern so accuracy compounds.

Build Something Like This More Case Studies

Available for new work

Have a backend, AI, or data problem worth solving?

From production APIs to self-hosted AI that kills per-call costs let's scope it. I reply within one business day.

Start a Project Connect on LinkedIn