Trust Intelligence Methodology
Deterministic evaluation algorithms and multi-layered verification models.
The Core Engineering Question: "Can I trust this?"
DETERMINISTIC & EXPLAINABLE SECURITY
The Truth Engine replaces opaque binary classifications ("Safe" vs "Unsafe") with multi-layered, explainable risk profiling. By combining mathematical security ratings (SSL, WHOIS, VirusTotal) with natural language comprehension (Google Gemini AI), we formulate a 0-100 Trust Verdict backed by verifiable evidence trails.
Score Weight Allocation
Website & URL Vector (100% Weight Tally)
Validates transport layer encryption & certificate chain authority.
Inspects hostnames for brand spoofing, typosquatting & IDN homoglyphs.
Analyzes creation timestamp. Newly registered domains carry risk penalties.
Cross-checks with 70+ antiviruses and domain reputation databases.
Queries official Google phishing and malware registries.
Email Inspection Vector
Checks freemail tags and SPF/DKIM alignment indicators.
Recursively scans target URLs embedded in mail body layers.
Checks for coercion, urgency, financial fraud & phishing copy.
Images & QR Code Media
Checks editor signatures, camera hardware data & timestamps.
Decodes QR image payload and passes target URLs to verification cluster.
Visual AI models detect deepfake manipulation & document spoofing.
PDF & Text Documents
Checks PDF authoring software, incremental updates & JavaScript streams.
Verifies embedded anchors in text blocks and form fields.
Detects scam patterns, prompt injection, and invoice manipulation.