ARCHITECTURE & METHODOLOGY

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)

HTTPS & SSL Configuration15%

Validates transport layer encryption & certificate chain authority.

Phishing Brand Heuristics20%

Inspects hostnames for brand spoofing, typosquatting & IDN homoglyphs.

WHOIS Registration Age20%

Analyzes creation timestamp. Newly registered domains carry risk penalties.

VirusTotal Database25%

Cross-checks with 70+ antiviruses and domain reputation databases.

Google Safe Browsing20%

Queries official Google phishing and malware registries.

Email Inspection Vector

Sender Domain Match25%

Checks freemail tags and SPF/DKIM alignment indicators.

Body Hyperlink Extraction25%

Recursively scans target URLs embedded in mail body layers.

Gemini Semantic Audit50%

Checks for coercion, urgency, financial fraud & phishing copy.

Images & QR Code Media

EXIF Metadata Analysis30%

Checks editor signatures, camera hardware data & timestamps.

QR Payload Resolution20%

Decodes QR image payload and passes target URLs to verification cluster.

Gemini Vision & OCR50%

Visual AI models detect deepfake manipulation & document spoofing.

PDF & Text Documents

Structural Metadata25%

Checks PDF authoring software, incremental updates & JavaScript streams.

Embedded Link Scanning25%

Verifies embedded anchors in text blocks and form fields.

Gemini NLP Audit50%

Detects scam patterns, prompt injection, and invoice manipulation.