Evidence and risk methodology
Trustiry keeps identity evidence separate from automated risk observations and publishes the limits of each.
Last updated: 17 July 2026
Trustiry produces a versioned decision from evidence available at a stated time. A decision includes the normalized domain, relationship status, risk status, intended action, reason codes, unknowns, policy version, and expiry. It is not a certification or a safety guarantee.
What we measure
The two numeric fields answer different questions:
- Reviewed relationship score: shown only when current evidence accepted by a reviewer connects the domain to the named organization. Traffic, HTTPS, age, and reputation do not create it.
- Observed risk score: a weighted total of warning signals seen during the current check. Zero means no weighted warning was observed in available sources, not that a site is safe.
The signals behind them
| Signal | What it tells us |
|---|---|
| Traffic and web presence | Used to prioritize crawling and as display context only. It never establishes identity or safety. |
| Connection security | HTTPS is reported as transport information. Missing HTTPS adds risk, but HTTPS itself adds no relationship score. |
| Lookalike & impersonation patterns | Whether a domain mimics a well-known brand (character swaps, added words, suspicious endings). |
| Threat intelligence | Matches against phishing, malware, and newly-registered-domain signals. |
| Official relationship | Time-bounded evidence accepted after domain-control proof and two-person administrative review. |
| Community reports | Only reviewed reports with a recorded outcome can affect public risk signals. |
Each link-check response includes its signal breakdown and scoring-model version. Informational observations carry zero points. Risk weights are capped at 100 and can change only with a new model version.
How we decide something is “official”
A claimant first proves domain control with a time-limited DNS TXT record. That does not prove the claimant represents the named organization. Relationship evidence is reviewed separately, and two different authorized reviewers must accept it before the public marker appears. Evidence can expire, be withdrawn, or be disputed.
Confidence, and when we hold back
Confidence describes evidence coverage, not the chance that a site is safe. Missing domain age, missing threat-feed coverage, and an unconfirmed organization relationship appear as unknowns. Sensitive actions fail closed when current relationship evidence is absent. Public accusations and feed entries require review rather than a raw automated check.
We can be wrong, here’s your recourse
If a verdict about your site is inaccurate, or you want to claim it, you can dispute it: use the “this is wrong / claim this site” option on the profile, or email legal@trustiry.com. We review disputes, fix genuine errors, and can suppress a contested label while we investigate. More detail is in our Acceptable Use & DMCA policy.
Our data sources
Trustiry is built on open and licensed data, including large-scale web-crawl and ranking datasets, IP-geolocation data (DB-IP and open sources), public threat signals, publisher feeds for content, and reports from our community, all combined in our own graph rather than resold from a single vendor.