FlagWise v1.0.0 - Shadow AI Detection System

Major Features

Core Functionality

• Real-time LLM Traffic Monitoring - Captures and analyzes outgoing LLM requests
• Shadow AI Detection - Identifies unauthorized AI usage across your network
• Risk Assessment Engine - Scores requests based on content and behavioral patterns
• Multi-Provider Support - Monitors OpenAI, Anthropic, Cohere, Google, and HuggingFace traffic

Security & Compliance

• Detection Rules Engine - Configurable keyword, regex, and model restriction rules
• Real-time Alerting - Immediate notifications for high-risk activities
• User Activity Tracking - Monitor individual user LLM usage patterns
• Audit Trail - Complete logging of all LLM interactions

Analytics & Reporting

• Interactive Dashboard - Real-time analytics with volume trends and threat detection
• Provider Analytics - Usage breakdown by AI provider and model
• Risk Scoring - Comprehensive threat assessment metrics
• Historical Analysis - Track patterns over time with multiple time ranges

Technical Architecture

• Kafka-based Streaming - High-throughput message processing
• PostgreSQL Database - Robust data persistence with analytics aggregation
• Docker Containerization - Easy deployment and scaling
• REST API - Comprehensive API for integrations
• React Dashboard - Modern, responsive web interface

What's New in v1.0.0

Enhanced User Experience

• Improved Search Functionality - Search works when live feed is paused
• Better Chart Visualization - 13px fonts for improved readability
• Inline Schema Editing - Direct Kafka schema editing without modal dialogs
• Realistic Data Models - Focus on network-capturable data only

System Optimizations

• Removed Forecasting Complexity - Simplified analytics for better performance
• Token-Free Architecture - Accurate network monitoring without response data
• Cleaner Codebase - Removed 368+ lines of unnecessary complexity
• Enhanced Data Generation - More realistic traffic simulation

Production Readiness

• Complete Documentation - API docs, setup guides, and usage examples
• Health Monitoring - System health endpoints and status indicators
• Role-based Access - Admin and viewer user permissions
• Security Hardening - JWT authentication and input validation

Quick Start

Docker Deployment

git clone https://github.com/bluewave-labs/flagwise.git
cd flagwise
docker-compose up -d

Access Points

• Dashboard: http://localhost:3000
• API Documentation: http://localhost:8000/docs
• Health Check: http://localhost:8000/health

Default Credentials

• Username: admin
• Password: admin123

System Requirements

• Docker & Docker Compose
• 4GB+ RAM recommended
• Network access for LLM provider monitoring

Security Features

• Content Analysis - Detects sensitive information in prompts
• Behavioral Monitoring - Identifies unusual usage patterns
• Model Restrictions - Block specific AI models or providers
• Risk Scoring - Automated threat assessment (0-100 scale)
• Alert Management - Configurable notification thresholds

Use Cases

• Enterprise IT Security - Monitor employee AI tool usage
• Compliance Monitoring - Ensure AI usage meets regulatory requirements
• Shadow IT Detection - Discover unauthorized AI service adoption
• Risk Management - Identify potential data leakage through AI prompts
• Usage Analytics - Understand AI adoption patterns across organization

Integration Ready

• REST API - Full programmatic access to all functionality
• Webhook Support - Real-time notifications to external systems
• Export Capabilities - Data export for compliance and reporting
• Custom Rules - Configurable detection logic for your environment

This release represents a production-ready shadow AI detection system designed for enterprise security teams, compliance officers, and IT administrators who need visibility into AI usage across their organization.