What it checks

Signal Description
Email reputation Flags disposable domains, shared role accounts, and public email providers
Message content Classifies the message for spam keywords, SEO/backlink pitches, excessive links, HTML, all-caps, repeated punctuation, and low-context length
Malicious links Extracts URLs from the message and the website field and checks them against Verifence threat sources
Language Detects the language of the message

How scoring works

Each signal contributes to a risk score from 0–100. The score maps to a verdict:

Verdict Risk score Meaning
clean 0–24 No meaningful risk signals
review 25–79 Some risk signals — worth a human look
spam 80–100 Strong risk signals

A known-malicious URL forces the score to 100. Reputation-provider failures are treated as non-fatal — they never turn a legitimate lead into spam.

How to analyze

Submit a lead with POST /api/lead/analyze. Either email or message must be provided; name, company, website, and source are optional.

Request
{
  "email": "jane@example.com",
  "message": "Interested in a demo for our team.",
  "name": "Jane Doe",
  "company": "Example Inc",
  "website": "https://example.com",
  "source": "contact-form"
}
Response
{
  "request_id": "3f1c...uuid",
  "verdict": "clean",
  "risk_score": 5,
  "categories": ["public_email"],
  "reasons": ["The email uses a public email provider."],
  "language": "en"
}

Result fields

  • ✓request_id — Unique ID for the analysis, stored for later reference
  • ✓verdict — clean, review, or spam
  • ✓risk_score — Integer from 0 to 100
  • ✓categories — Machine-readable signal categories (e.g. disposable_email, seo_pitch, malicious_url)
  • ✓reasons — Human-readable explanations for each signal
  • ✓language — Detected message language, or null if no message was provided

Examples

  • ✓Score contact-form submissions and auto-route spam verdicts away from sales.
  • ✓Reject signups using disposable email domains while letting business addresses through.
  • ✓Flag unsolicited SEO and backlink outreach before it reaches your inbox.
  • ✓Block leads whose message or website links to known malicious infrastructure.