Prerequisites
Hardware requirements
- CPU: 4+ cores (8+ recommended for multi-stream)
- RAM: 8GB minimum (16GB+ recommended for 8 streams)
- Storage: 20GB (for GGUF models and database)
- Network: Gigabit Ethernet (for RTSP streaming)
Software requirements
- Python: 3.11 or higher
- Node.js: 18 or higher
- llama.cpp: latest version with server support
Installation
1. Clone the repository
git clone <repository-url>
cd AEGIS2
2. Backend setup
cd backend
pip install -r requirements.txt
3. Install llama.cpp and models
# Install the llama.cpp server
# Download from https://github.com/ggerganov/llama.cpp/releases
# Start llama.cpp server with the Ministral-3 model
llama-server -hf mistralai/Ministral-3-8B-Instruct-2512-GGUF:Q4_K_M --port 8080 -ngl 99
4. Frontend setup
cd ..
npm install
Running the application
Start the backend server
python backend/server.py
# Server starts at http://localhost:8000
Start the frontend dev server
npm run dev
# Frontend available at http://localhost:5173
Usage guide
- Navigate to the Multi-Stream tab: click "Multi-Stream CCTV" in the top navigation.
- Add a stream: click "+ Add Stream" and fill in details (RTSP URL, webcam index, or file).
- View live streams: monitor up to 8 streams simultaneously in the grid view.