AI-Powered · Real-Time · Fully Offline

Detect Fire
In Seconds

Vigil is an AI-powered fire and smoke detection system that integrates with your existing IP cameras — no additional hardware required.

Vigil — Warehouse Camera 3
🔥
CAM-03 · 25 FPS · GPU FIRE DETECTED
<2s
Average detection time
16+
Simultaneous cameras
24/7
Continuous monitoring
0
Internet required
Features

Every Detail Considered

Designed for industrial environments — easy to install, minimal maintenance, works with the cameras you already have.

🔥
Real-Time Fire Detection

Advanced AI models classify flames and smoke within milliseconds. High accuracy keeps false alarm rates minimal.

🎥
Multi-Camera Support

Connects to your existing IP cameras via RTSP. Monitors 16+ cameras simultaneously — no extra hardware needed.

🔔
Instant Notifications

Sends automatic email, SMS, and Telegram alerts the moment fire is detected. Alarm snapshot attached automatically.

GPU Acceleration

NVIDIA CUDA and Apple MPS support dramatically reduces processing time. Also runs on CPU when GPU is unavailable.

🗂️
Alarm History & Snapshots

Every detection is saved with an annotated image and timestamp. Full incident archive available at a glance.

🔒
Fully Offline

No internet connection required. All AI processing runs on local hardware — your critical data never leaves the premises.

How It Works

From Setup to Alert in 4 Steps

No complex installation. Add your existing cameras and Vigil handles the rest.

1
Add Your Cameras

Enter each camera's IP address and RTSP stream details. Username/password authentication supported.

2
AI Model Loads Automatically

Vigil loads the AI model and selects GPU or CPU automatically. No additional configuration required.

3
Continuous Monitoring Begins

Inference rate increases when motion is detected; stays low in idle environments to minimise resource usage.

4
Alarm & Notification

On fire detection: audible alert, on-screen notification, and automatic messages to configured channels.

🎥
IP Camera StreamRTSP · H.264 · H.265
LIVE
👁️
Motion AnalysisFrame difference · Adaptive threshold
FILTER
🧠
AI AnalysisGPU batch processing · <50ms latency
AI
🔥
Fire / Smoke DetectionBounding box · Confidence score
ALARM
🔔
Notifications SentEmail · SMS · Telegram · Audio
SENT
Technical Specifications

Technical Details

AI & Model
Model ArchitectureDeep Learning (Object Detection)
Detection ClassesFire · Smoke
Confidence Threshold45% default (adjustable)
GPU SupportCUDA · Apple MPS · CPU
Batch Inference8 frames / pass
Platform & Requirements
Operating SystemWindows 10/11 · macOS 11+
Minimum RAM8 GB
Recommended GPUNVIDIA RTX Series
Internet RequiredNo
InstallationOne-click setup (.exe)
Camera & Connectivity
ProtocolRTSP (TCP / UDP)
Simultaneous Cameras16+ (hardware dependent)
Motion SensitivityLow · Medium · High
On DisconnectAuto reconnect
Video CodecH.264 · H.265
Notifications & Logging
EmailSMTP · TLS · SSL
SMSAPI integration
TelegramBot + image delivery
Snapshot StorageJPG · Date/time folders
Alarm History2,000 records (local)

Protect Your Facility Today

Request a demo or get a custom quote for your project.