Sentinal Evidence
An AI-Powered Cybersecurity Platform with Local AI
Sentinal is an AI-powered cybersecurity platform designed to detect security threats locally, without sending sensitive information to cloud-based AI services.
During a security threat, sensitive files and system information may be exposed. Although cloud AI can provide powerful and fast analysis, sending this information to an external server introduces potential privacy and data-leakage risks.
Our Solution
To address this, we implemented Local AI, where threat detection and analysis happen directly on the user’s computer.
Our system uses Qwen3:4B, running locally through Ollama, allowing the AI to analyze threats without requiring sensitive data to leave the device.
How We Built It
- Frontend: HTML, CSS, React, and JavaScript
- Backend: Python
- API: FastAPI, which connects the frontend with our Python backend
- Local AI: Qwen3:4B running through Ollama
- Architecture: Threat detection and AI analysis are performed locally on the user’s system.
Future Improvements
We plan to further improve Sentinal by implementing:
- More capable local AI models for better threat detection and analysis
- Multi-device deployment to protect multiple systems from a centralized platform
- Firewall integration to automatically block or respond to detected threats
- Real-time monitoring and automated threat response
- Improved threat intelligence for identifying new and evolving attack patterns
Our Experience
Building Sentinal was a valuable hands-on experience for our team. Instead of relying heavily on AI to build the project for us, we focused on understanding the workflow, architecture, integration, and development process ourselves.
This project helped us understand how different technologies work together to build a practical cybersecurity solution—from the frontend and backend to APIs and locally deployed AI.
Team Root Access
Top comments (0)