SENTINEL
Open-source behavioral intelligence platform for detecting child groom
SENTINEL is an open-source behavioral intelligence platform for detecting child grooming on digital platforms. Released by Sentinel Foundation in April 2026, it analyzes patterns of user behavior over time rather than scanning message content for keywords, identifying grooming sequences before direct harm occurs.
Key capabilities:
- Behavioral scoring across four signal types: linguistic register shifts, relationship graph evolution, temporal escalation patterns, and fairness gates
- Tamper-evident audit logs (cryptographically chained, 7-year retention) for legal proceedings and regulatory compliance
- Privacy-preserving federation: platforms share encrypted threat signatures without exposing raw messages or user identities
- Demographic parity audits enforced as a deployment gate before any detection model goes live
- Built-in compliance for EU Digital Services Act, UK Online Safety Act, COPPA, and GDPR
Free for organizations under $100k annual revenue and all non-commerci