Interpret
AI reads text, image, or voice and identifies observable clues.
PrivacyBack to NIGRAH INSIDE THE INTELLIGENCE SYSTEM
From one clue
to a connected pattern.
NIGRAH brings AI-assisted sample analysis together with a structured financial threat taxonomy and evidence-linked intelligence. Each layer adds context to what came before.
25warning signs
10behavior patterns
15threat families
Map
NIGRAH links those clues to defined warning signs and manipulation patterns.
Score
The NIGRAH scoring rules weight unique signals and measure pattern and family overlap.
Ground
Reviewed public evidence provides source quality and supports a complete Threat Index.
THE DIVISION OF WORK
AI interprets.
NIGRAH structures the result.
The AI layer reads the submitted sample and proposes observable signals with the text or visual cue that supports each one. NIGRAH maps those observations to its versioned taxonomy, checks the linked pattern structure, and calculates the scores using its own rules.
That separation makes each result easier to inspect: the observation, its taxonomy connection, and its score are shown as distinct steps.
HOW CLUES CONNECT
A warning sign can point to
several possible behaviors.
Select a warning sign and follow its links through the catalog. Shared patterns can connect different scam families.
FOLLOW A RED FLAG
Pick a warning sign. See where it leads.
Warning signs can point to tricks scammers use; those tricks can appear in different scam types. Choose one clue to explore the catalog links.
2 related tricks · 5 scam types
These links come from the reviewed model catalog. They show possible relationships, not live reports or proof about a specific message.
SCORING FRAMEWORK · VERSION 1.2
Each score answers
a different question.
Scores are built from identified signals, catalog relationships, and linked public evidence. The components stay separate so users can see how each part contributes.
TSS
Threat Signal ScoreAdds the fixed weights of the unique warning signs identified in a sample.
TSSN
Normalized Threat Signal ScorePlaces the weighted signal total on a 0–100 scale for easier comparison.
Pattern score
How signals group togetherMeasures how strongly identified signals support activated manipulation patterns.
EQ
Evidence QualityAssesses the quality of reviewed public sources linked to the analysis.
CS
Confidence ScoreCombines evidence quality, signal strength, pattern consistency, and family stability.
TI
Threat IndexCombines the normalized signal score, pattern consistency, family harm, family similarity, and confidence.
A sample can show signal and pattern scores before a public evidence source is linked. Evidence Quality, Confidence Score, and Threat Index are calculated only when their required inputs are available.
READING THE SAMPLE RESULT
Color guides attention.
The color is a visual attention cue based on matched catalog signals. It does not determine the legal status or safety of a person, firm, or message.
FOLLOW THE METHOD
Explore a sample with NIGRAH.
See the observations, signal links, and score components together.