Stop accepting leads from URLs you can't trust.
AI Domain Scrub runs 50+ deterministic checks across six signal groups plus an AI-evaluated TCPA disclosure review, then cross-references every URL against an owner graph that catches sister sites — fresh-looking domains run by operators behind landing pages you have already rejected. Use any combination of those signals in your boberdoo buyer rules — or fall back to the blended trust score for the simple cases.
50+ checks per URL · 21 identity fingerprints · Sister-site detection · <200ms p95 on cached domains
Bad URLs cost you twice.
When a lead lands on a misconfigured or non-compliant page, the cost lands on the buyer.
When a lead's landing page is misconfigured, expired, or quietly running a TCPA-non-compliant disclosure, the cost lands on the buyer. You pay for the lead, you absorb the contact-rate hit, and you carry the compliance exposure if the consent record can't hold up.
Visual inspection doesn't scale to the volume boberdoo customers move. AI Domain Scrub does the inspection for you — every URL, every time, before the lead enters your funnel.
Six signal groups. Fifty-plus checks. All available to your buyer rules.
Use any combination of these signals — or all of them — in your boberdoo outside-service rules. Every response includes the full per-check detail so you can pick the data points that matter to your business.
Domain & WHOIS
Domain age, registrar reputation, TLD risk, recent transfers, WHOIS privacy posture.
DNS & Hosting
IP geolocation, ASN reputation, email authentication (SPF / DKIM), reverse DNS.
SSL & HTTP
HTTPS enforcement, certificate validity and issuer, security headers, redirect chain.
Content & Identity
Privacy policy, terms, contact info, physical address, anonymity-state flag, tracking pixels.
Dark Patterns
Fake urgency, incentivized clicks, pre-checked consent, false endorsement, bait-and-switch.
TCPA Disclosure (AI-evaluated)
Disclosure proximity, contrast, font size, opt-in scope, completeness — graded by a purpose-built evaluator.
Catch the sister site before it catches you.
Fraudsters rarely run one landing page. They run dozens — sometimes hundreds — under fresh-looking domains, all generating leads back to the same operation.
Each domain looks different on its face: a different name, a different design, a different vertical. But they share fingerprints. The same Google Analytics ID. The same favicon hash. The same schema.org Organization name. The same footer copyright string. The same Tag Manager container.
AI Domain Scrub captures up to 21 identity fragments from every URL it scores and uses them to build a cross-customer owner graph. When a fresh, good-looking domain comes through your funnel and clusters with domains you have already rejected — or that another buyer in the boberdoo network has rejected — the response flags the relationship before you accept the lead.
You get back the related domains, the exact shared fingerprints, and a cluster risk flag. Your buyer rule can downgrade an apparent `approve` to `review` whenever cluster risk is high, or reject outright on a confirmed sister-site match.
“A 22-year-old domain with a clean WHOIS can still be one of three sites a single operator stood up last quarter. AI Domain Scrub shows you the cluster.”
Or skip the rule-building and use the AI verdict.
If you'd rather not weigh signals individually, every response also includes a blended 0–100 trust score and an approve / review / reject recommendation. Map the recommendation to your boberdoo rule once and you're done.
Approve · Score 80–100
Established domain, clean security posture, compliant disclosure. Accept the lead.
Review · Score 60–79
Mixed signals — usually one or two meaningful gaps that warrant a human check before accepting at scale.
Reject · Score 0–59
Critical failure (no SSL, malware blacklist, missing privacy policy) or low score. Do not accept.
How AI Domain Scrub fits into your outside service.
A single HTTPS POST in. The full per-check breakdown out — including owner-cluster risk.
Cached scores return in under 200ms (p95). First-time domains return a queued response immediately and finish scoring within ~60 seconds (p95) — your outside service re-polls, no synchronous wait.
Four-step flow, end to end
Your boberdoo outside service POSTs the URL. AI Domain Scrub runs the checks and the owner-graph lookup. Your buyer rule applies whichever signals matter to you.
Four steps from URL to decision.
Wire the signals you care about — or the AI recommendation — into your existing boberdoo rule. No score parsing, no substring matching.
Buyer outside service POSTs the lead URL.
Your boberdoo outside service sends a single HTTPS POST to AI Domain Scrub with the landing page URL and a small context payload (affiliate ID, buyer ID, source IP).
AI Domain Scrub runs the checks.
50+ deterministic checks across six signal groups, plus an AI evaluator on the TCPA disclosure, plus a cross-customer owner-graph lookup that catches sister-site clusters.
AI Domain Scrub returns the full picture.
Per-check pass / warn / fail / info / skip detail, the owner-cluster risk flag with related domains and shared fingerprints, and — for buyers who want it — a blended 0–100 trust score with an approve / review / reject recommendation. All in one JSON response.
Your buyer rule applies it.
Wire whichever fields matter to your business — individual checks, the cluster risk flag, the AI recommendation, or any combination — into your boberdoo outside-service rule.
What we store. What we don't.
AI Domain Scrub scores the destination URL, not the lead.
What we store
A short audit trail and the score itself. Nothing about the lead.
- Domain-level score and check results, with a 30-day TTL.
- Per-call audit log: timestamp, buyer ID, domain, score, duration. PII fields in the context payload are SHA-256 hashed before persist. 1-year TTL.
- Identity fragments for the owner graph — hashes by default; raw values only for non-PII signals like analytics IDs.
What we don't store
Nothing about the consumer behind the lead.
- No raw lead PII from your context payload.
- No request bodies past the context blob.
- No browser fingerprints. No end-user tracking.
Common questions
By eTLD+1. `https://www.example.com/path?q=1` and `https://example.com/another` both resolve to `example.com`.
Under 200ms (p95) for cached domains. First-time domains return a queued response immediately and finish scoring within ~60 seconds (p95). Your outside service re-polls — there's no synchronous wait.
Already-cached domains keep serving at full speed. Cold-cache requests return a `scoring_unavailable` response with a retry hint instead of failing the call. These responses are operational and aren't billed.
No. Boberdoo billing only counts a call when AI Domain Scrub returns a numeric score.
AI Domain Scrub is currently invite-only and provisioned per buyer through your account manager. Each buyer gets a key tied to its own usage plan, with rate limits and quota separate from every other buyer.
Yes. Every response includes the full per-check breakdown — pass / warn / fail / info / skip with a one-line `detail` for each. The same response includes the AI evaluator's red flags, green flags, and a short reasoning paragraph for borderline cases.
Ready to plug AI Domain Scrub into your outside service?
AI Domain Scrub is rolling out to boberdoo customers in phases. Your account manager will issue an API key, walk you through the outside-service configuration, and run a smoke test against a known-good and a known-bad URL before you go live.
About AI Domain Scrub
AI Domain Scrub is operated by boberdoo for boberdoo lead buyers. Self-serve signup is not available. API keys are issued by your account manager.