Open-source AI security frameworks

Open frameworks for
practical AI security defenses

AIDEFEND is available today through the framework, the MCP service, and advisory support. FraudDEFEND is inspired by the MITRE F3 framework and focused on fraud defense.

Explore AIDEFEND See AIDEFEND in Action
OPEN BY DESIGN
Open Frameworks

Frameworks from AIDEFEND Labs

AIDEFEND is available today through the framework, the MCP service, and advisory support. FraudDEFEND is inspired by the MITRE F3 framework and focused on fraud defense.

Live Now Flagship

AIDEFEND

AIDEFEND is available today through three core offerings: the framework, the MCP service, and advisory support for teams that want help putting the work into practice.

  • Framework: the open foundation for AIDEFEND's AI security work.
  • MCP Service: a practical service layer connected to the framework.
  • Advisory: consulting and implementation support for teams adopting the work.
Visit AIDEFEND MCP Service Advisory Support
Coming Soon Next Framework

FraudDEFEND

FraudDEFEND is inspired by the MITRE F3 framework and focused on fraud defense.

Why Open

Why publish the frameworks in public

AI security knowledge should be easier to review, use, and build on. Publishing the frameworks in public helps more teams turn defensive ideas into real implementation.

MITRE ATLAS MAESTRO OWASP LLM Top 10 OWASP Agentic Top 10 OWASP ML Top 10 Google SAIF 2.0 NIST AML 2025 Cisco AI Security Framework Databricks DASF 3.0