The model powering our agents.
MF-1 is the foundation of MindFort's AI Security Agents. A custom LLM purpose-built for offensive security — trained on vulnerability research, exploit development, and penetration testing. This is the core technology we're bringing to every security domain.
Security-first, not security-adapted.
General-purpose models are trained to avoid security topics. MF-1 is the opposite — trained specifically on offensive security. It generates working exploits, reasons about attack chains, and powers our agents to find vulnerabilities that matter.
API Access Available
Direct access to MF-1 for your own security tooling
MF-1 Architecture
Transformer Model
"analyze JWT auth bypass vectors"
{"exploit": "alg:none", "confidence": 0.94}
120B
Parameters
120
Tokens/sec
v2.4
Latest Release
AppSec today. Every security domain tomorrow.
We're building agents for continuous pen testing now. Our goal: deploy an entire security organization of specialized agents for every domain.
Continuous pen testing with AI Security Agents
Security testing on every deploy, automatically
Specialized agents for API security testing
Agents for cloud infrastructure and configuration
Agents for every security domain — network, identity, compliance