Vigilant is a proactive, AI-driven system monitoring platform designed to analyze Windows system health in real time and detect potential risks before they lead to crashes or data loss. It combines a lightweight background agent, centralized log processing, and AI-powered anomaly detection to provide actionable insights, risk assessments, and preventive recommendations.
- Overview
- Why Vigilant?
- System Architecture
- How It Works
- Dashboard Insights
- Alert System
- Analysis Report Generation
- AI Model
- Tech Stack
- Security Features
- Key Capabilities
- Use Cases
- Installation
Traditional system monitoring tools show raw logs and metrics. Vigilant goes further by:
- Continuously collecting system events and hardware metrics
- Analyzing the last 2 hours of activity using AI (Perplexity Sonar-Pro)
- Detecting hardware degradation, credential issues, privilege anomalies
- Estimating crash probability and performance impact
- Generating structured analysis reports
- Notifying users instantly via email when critical issues occur
Vigilant focuses on prevention, not just monitoring.
Problem: System failures often occur without warning, leading to data loss, downtime, and productivity impact.
Solution: Vigilant provides early warning signals by analyzing system behavior patterns, identifying anomalies, and recommending preventive actions before failures occur.
Key Differentiators:
- AI-powered root cause analysis
- Proactive crash probability estimation
- Automated alert system with context
- Actionable recommendations with priority levels
- Real-time correlation of system events and hardware metrics
Vigilant follows a distributed architecture:
User → Web Platform → Download Agent → Background Monitoring
↓
Windows Agent (vigilant-agent.exe)
↓
Collects logs every 1 minute
↓
Backend API (Node.js + Express)
↓
MongoDB (Event + Metrics Storage)
↓
AI Analysis (Perplexity Sonar-Pro)
↓
Insights + Risk Assessment + Alerts
Frontend (Web Dashboard)
- User authentication and session management
- Real-time system status visualization
- Analysis report generation interface
- Alert configuration panel
Backend API
- RESTful endpoints for agent communication
- JWT-based authentication layer
- Log aggregation and preprocessing
- AI analysis orchestration
- Email notification service
Monitoring Agent
- Lightweight Windows background process
- Event log extraction (System, Hardware, Security)
- Performance metrics collection (CPU, Memory, Disk)
- 1-minute polling interval
- Secure API communication
Database Layer
- MongoDB for event storage
- Time-series optimized schema
- Efficient querying for 2-hour analysis windows
AI Analysis Engine
- Perplexity Sonar-Pro integration
- Pattern recognition in system logs
- Risk scoring algorithm
- Recommendation generation
Users log in securely using JWT-based authentication. After login, users download the vigilant-agent.exe executable.
The Vigilant Agent:
- Runs as a lightweight Windows background process
- Sends system events and metrics every 1 minute
- Collects hardware/system/boot events, security audit logs, credential manager events, privilege assignment logs
- Monitors CPU, memory, and disk metrics
- Can store logs locally if internet is unavailable (optional enhancement)
- Automatically resumes transmission once connectivity is restored
The backend:
- Stores logs in MongoDB with timestamp indexing
- Aggregates the most recent 2 hours of activity
- Prepares structured log summaries for AI analysis
- Maintains data retention policies
Vigilant uses Perplexity Sonar-Pro to analyze system activity. The AI evaluates:
- Recurring hardware errors and Machine Check Exceptions
- Credential cache failures
- Privilege escalation frequency
- Unusual login activity patterns
- Disk I/O and performance degradation
- Crash indicators and BSOD precursors
After analysis, Vigilant provides:
- Operational state monitoring
- Monitoring required indicators
- Hardware health signals
- Crash Probability: LOW / MEDIUM / HIGH classification
- Performance Impact: Quantified degradation metrics
- Data Integrity Status: Risk to data consistency
Examples:
- Recurring correctable hardware errors
- Credential Manager cache misses
- High frequency of user privilege checks
- Abnormal boot sequence patterns
Each action includes:
- Priority level (High / Medium / Low)
- Estimated time to resolve
- Difficulty level
- Clear step-by-step instructions
AI-generated suggestions such as:
- Monitor specific Event ID frequency thresholds
- Watch for BSOD indicators
- Track disk errors alongside hardware failures
- Observe memory allocation patterns
When warning or critical events are detected:
- Users receive automated email notifications
- Alerts are triggered for:
- High-frequency hardware errors
- Crash probability spikes
- Security anomaly patterns
- Critical system event sequences
- Users can generate a structured analysis report for deeper investigation
- Alert configuration allows customization of thresholds
Users can generate a detailed AI-based system analysis report including:
- Time window analyzed
- Event summaries and frequency distributions
- Risk classification with confidence scores
- Root-cause explanations
- Recommended actions with implementation details
- Prevention tips and best practices
This helps users:
- Understand what is happening in their system
- Take corrective action early
- Prevent system failure before it occurs
- Document system behavior for troubleshooting
Perplexity Sonar-Pro
- Optimized for contextual log interpretation
- Converts raw system events into human-readable insights
- Pattern recognition across temporal sequences
- Contextual understanding of Windows event relationships
- React.js - Component-based UI framework
- TailwindCSS - Utility-first styling
- Responsive Dashboard UI
- Node.js - Runtime environment
- Express.js - Web application framework
- RESTful APIs
- JWT Authentication
- MongoDB - Document-oriented storage
- Time-series optimized collections
- Perplexity Sonar-Pro API
- Windows executable (
vigilant-agent.exe) - Event log extraction via Windows APIs
- Metric collection at 1-minute intervals
- Secure HTTPS communication
- Email-based alert system
- Template-driven notification content
- JWT-based authentication: Secure token-based session management
- Machine ID verification: Agent identity validation
- Secure API communication: HTTPS-only data transmission
- Role-based access: Extendable permission system
- Data encryption: At-rest and in-transit encryption
- Audit logging: Comprehensive activity tracking
- Real-time monitoring with 1-minute granularity
- AI-driven anomaly detection
- Crash probability estimation
- Automated alert system with context
- Intelligent recommended actions
- Prevention-focused insights
- Structured report generation
- Historical event correlation
- Developers: Proactive crash prevention during development cycles
- Power Users: Monitoring system stability for critical workloads
- Security Teams: Tracking privilege anomalies and credential issues
- IT Professionals: Hardware degradation detection and replacement planning
- System Administrators: Early-stage system reliability monitoring
- Node.js (v14 or higher)
- MongoDB (v4.4 or higher)
- Windows OS (for agent deployment)
# Clone the repository
git clone https://github.com/yourusername/vigilant.git
cd vigilant
# Install backend dependencies
cd backend
npm install
# Configure environment variables
cp .env.example .env (message me for details)
# Edit .env with your configuration
# Start the backend server
npm start# Navigate to frontend directory
cd frontend
npm install
# Start development server
npm run dev- Log in to the web dashboard
- Navigate to the "Download Agent" section
- Download
vigilant-agent.exe - Run the agent with administrator privileges
- Agent will automatically register with your account
Mail: maheshn0418@gmail.com Project Link: https://vigilant-cyberx.vercel.app/
Vigilant is not just a monitoring tool — it is a proactive AI-driven system intelligence platform designed to help users detect and resolve system risks before failure occurs.