Full Transparency

How We Score Brand Trust

Every trust score on RegisteredBrands.AI is calculated using the same methodology, applied consistently to every brand. No exceptions. No pay-to-play. Here's exactly how it works.

Core Principles

Consistent Methodology

Every brand is scored using the same 7 checks with the same weights. No brand receives preferential treatment. The same rules apply whether you're a Fortune 500 or a startup.

Verified Data Only

Self-reported data earns zero points. Points are only awarded when our verification engine confirms credentials against authoritative sources (USPTO, GLEIF, DNS, SAM.gov).

No Pay-to-Play

Brands cannot pay to improve their trust score. Paid plans unlock more capsules and API calls — never a higher score. The score reflects verified reality, not budget.

Score Range: 0 – 100,000

The maximum possible score is 100,000 points. Here's what each range means.

80,000 – 100,000
Sovereign

Fully verified across all dimensions. Trademark confirmed, domain ownership proven, business entity validated, and all security signals present.

60,000 – 79,999
Verified

Strong verification with most signals confirmed. May be missing one non-critical verification (e.g., EIN for non-US entities).

30,000 – 59,999
Established

Partial verification with key signals confirmed. The brand has demonstrated legitimacy but hasn't completed all verification steps.

1 – 29,999
Unverified

Minimal or no verification completed. The brand has registered but hasn't provided verifiable credentials.

0
Flagged

The brand has been flagged by the community or failed critical verification checks. Under review.

The 7 Verification Checks

Every brand capsule is evaluated against these 7 independent checks. Each check queries authoritative, public data sources.

Check 1: DNS TXT Verification

15,000 pts

Proves domain ownership by requiring a specific DNS TXT record. This is the gold standard for domain control verification — the same method used by Google, Microsoft, and certificate authorities.

Data Sources

  • DNS TXT record lookup via authoritative nameservers

How It Works

We generate a unique token for each brand. The brand owner adds a TXT record to their domain's DNS configuration. Our engine queries the domain's DNS records and confirms the token is present.

Scoring

15,000 points if the token is found. 0 points if missing or incorrect. No partial credit.

Check 2: USPTO Trademark Cross-Reference

18,000 pts

Verifies the brand's trademark registration with the United States Patent and Trademark Office. A registered trademark is one of the strongest signals of brand legitimacy.

Data Sources

  • USPTO TSDR (Trademark Status & Document Retrieval) API
  • USPTO TESS (Trademark Electronic Search System)

How It Works

We search the USPTO database for the brand name and trademark number provided. We verify the trademark is active (not expired, cancelled, or abandoned) and matches the registrant information.

Scoring

18,000 points for a verified, active trademark. 0 points if the trademark number is provided but cannot be verified. Self-reported trademark numbers earn no points until confirmed.

Check 3: GLEIF / LEI Business Entity Lookup

16,000 pts

Confirms the brand is a registered legal entity using the Global Legal Entity Identifier Foundation (GLEIF) database and OpenCorporates registry.

Data Sources

  • GLEIF LEI API (Legal Entity Identifier)
  • OpenCorporates API

How It Works

We search for the business entity using the provided DUNS number, LEI, or company name. We verify the entity is active and the registration details match the brand capsule information.

Scoring

16,000 points for a verified business entity. 0 points if the identifier is provided but cannot be verified. Self-reported business identifiers earn no points until confirmed.

Check 4: Domain Signals (SSL / DMARC / SPF)

13,000 pts

Analyzes the brand's domain security posture by checking SSL certificate validity, DMARC email authentication policy, and SPF records. These signals indicate a professionally managed domain.

Data Sources

  • SSL certificate chain analysis
  • DNS DMARC record lookup
  • DNS SPF record lookup

How It Works

We connect to the brand's domain to verify SSL certificate validity, issuer, and expiration. We query DNS for DMARC and SPF records that protect against email spoofing and phishing.

Scoring

13,000 points for valid SSL + DMARC + SPF. Partial credit is not awarded — all three must be present and correctly configured.

Check 5: EIN / SAM.gov Verification

10,000 pts

Cross-references the brand's Employer Identification Number (EIN) with the SAM.gov federal contractor database and SEC EDGAR filings.

Data Sources

  • SAM.gov (System for Award Management)
  • SEC EDGAR (Electronic Data Gathering, Analysis, and Retrieval)

How It Works

We search SAM.gov and SEC EDGAR for the provided EIN. A match confirms the entity is registered with the federal government, which is a strong legitimacy signal.

Scoring

10,000 points for a verified EIN match. 0 points if the EIN is provided but cannot be verified.

Check 6: Google Safe Browsing Check

Pass/Fail pts

Checks the brand's domain against Google's Safe Browsing database, which tracks known phishing sites, malware distributors, and deceptive websites.

Data Sources

  • Google Safe Browsing API v4

How It Works

We query Google's Safe Browsing API with the brand's domain. If the domain is flagged for phishing, malware, social engineering, or unwanted software, the brand fails this check.

Scoring

This is a pass/fail gate. A flagged domain triggers an automatic review and potential denial. No points are awarded — this check prevents fraudulent brands from receiving any trust score.

Check 7: AI Synthesis & Fraud Detection

Verdict pts

An AI model cross-references all verification signals, detects inconsistencies, and produces a final verdict with reasoning. This is the human-equivalent judgment layer.

Data Sources

  • All signals from checks 1-6
  • Cross-reference analysis
  • Pattern matching against known fraud vectors

How It Works

Our AI model receives all verification results and analyzes them holistically. It looks for inconsistencies (e.g., a brand claiming to be Apple but failing trademark verification), suspicious patterns, and red flags that individual checks might miss.

Scoring

The AI produces a verdict: auto_approved (high confidence, all signals align), auto_denied (clear fraud indicators), or flagged_for_review (mixed signals requiring human review).

Additional Score Components

Beyond the 7 automated checks, these factors also contribute to the final trust score.

Owner Identity (0 – 8,000 pts)

Partial credit for providing verifiable owner information: legal name (3,000), state of incorporation (2,500), business address (1,500), phone number (1,000). This is the only area where self-reported data earns points, because it's a verifiable claim tied to a real person.

Document Completeness (0 – 12,000 pts)

Points awarded for providing supporting documents (articles of incorporation, business licenses, etc.) that are analyzed by our AI. Each document can earn up to 3,000 points, capped at 12,000.

Community Reputation (starts at 10,000 pts)

Every brand starts with 10,000 community reputation points. Each community flag reduces this by 2,000 points. Three or more upheld flags trigger a 'flagged' status regardless of other scores.

Age / History Bonus (0 – 10,000 pts)

Longevity on the platform earns bonus points: 7+ days (500), 30+ days (2,000), 90+ days (4,000), 180+ days (7,000), 365+ days (10,000). This rewards brands that maintain their capsule over time.

Flag Penalty (–15,000 per upheld flag)

Each community flag that is upheld after review deducts 15,000 points from the total score. This is a significant penalty designed to protect the ecosystem from bad actors.

Minimum Score: 0

Trust scores cannot go negative. The minimum possible score is 0. The maximum possible score is 100,000. Scores are recalculated whenever new verification data becomes available.

About Unclaimed Profiles

Some brand profiles on RegisteredBrands.AI are generated from public data before the brand owner registers. Here's how we handle them.

What are unclaimed profiles?

To provide immediate value to AI agents querying brand trust, we pre-verify well-known brands using publicly available data. These profiles are clearly marked as "Unclaimed" and include a disclaimer: "Generated from public data. Not endorsed by [Brand]."

How are they different from claimed profiles?

Unclaimed profiles cannot earn DNS verification points (since the brand owner hasn't added the TXT record). They display no brand logos (text only). They carry a prominent amber "Unclaimed" badge. Their scores are typically lower because key verification steps require brand owner participation.

How can a brand claim their profile?

Brand owners can claim their profile by registering on the platform and completing the verification process. Once claimed, the profile is updated with the owner's information and can earn full verification points.

Dispute and removal

Brand owners can dispute their trust score by providing evidence via our dispute process. Brands can also request complete removal of their unclaimed profile — we honor all opt-out requests.

Questions About Your Score?

If you believe your brand's trust score is inaccurate, you can dispute it with evidence. We review every dispute within 48 hours.