AKI_Sentinel is a next-generation anti-cheat system for FiveM (ESX Legacy) that goes beyond traditional rule-based detection.
Instead of relying on static checks, it analyzes player behavior in real time, detects anomalies using machine learning, and classifies suspicious patterns into meaningful categories.
Tracks player activity such as event frequency, money flow, and item usage to detect abnormal patterns.
Identifies suspicious aiming behavior like unnatural headshot ratios and reaction times — fully server-side.
Uses anomaly detection (Isolation Forest) to identify behavior that deviates from normal player patterns.
Automatically categorizes suspicious behavior:
- Event Spam
- Money Exploit
- Item Duplication
- Aimbot / Aim Assist
Combines all signals into a dynamic risk score and determines actions:
- OK
- FLAG
- KICK
- BAN
Real-time web interface displaying:
- Player risk levels
- Detected cheat types
- System decisions
Every action is logged and traceable for full transparency and review.
- Lua (FiveM / ESX)
- Node.js (API + WebSocket)
- Python (Machine Learning)
- HTML/JS Dashboard
AKI_Sentinel does not assume players are cheating.
It evaluates behavior and answers one question:
“Is this statistically normal?”
- Fully server-side (no invasive client methods)
- Designed to minimize false positives
- Modular and extensible architecture
MIT License
</> BASH
''' cd backend npm install node server.js node api.js '''
</> BASH
''' cd ml python ml_engine.py '''
</> BASH
''' dashboard/index.html '''