A 4-stage analysis pipeline that classifies threats in under 200ms — from cache lookup to Cloud AI verdict
Every message passes through 4 stages. Most threats are caught before reaching the AI layer.
Previously analyzed messages are served instantly from cache. SHA-256 content hashing ensures identical messages get immediate verdicts without re-analysis.
PostgreSQL database of known scam hashes, malicious URLs, and flagged phone numbers. Community reports and automated feeds keep it current.
5,876+ regex patterns specifically designed for Ukrainian fraud. Covers Privat24, Monobank, Nova Poshta, Diia, military donation scams, and more.
Messages that pass all previous stages are analyzed by Cloud LLM (OpenAI/Claude/OpenRouter). Context-aware classification with explanation.
Each module handles a specific threat vector with dedicated algorithms and data sources
Unified endpoint for SMS, URLs, phone numbers. The core classification engine.
Proactive phishing domain detection via CT logs, DNS monitoring, and visual similarity scoring.
Multi-chain wallet screening, mixer detection, sanctions check. 11 blockchains supported.
@NoScamAIbot — free inline checks, group auto-moderation, forward message analysis.
Malware detection for files up to 50MB. Static analysis, signature scanning, VirusTotal integration.
Threat feeds, campaign clustering, entity graph, STIX 2.1 export for threat sharing.
Get your API key and start analyzing threats in minutes. Free tier includes 200 requests/day.
