How Warcast Works
Real-time global conflict intelligence powered by advanced AI orchestration
Our Mission
Warcast monitors global conflicts in real-time by aggregating news from 50+ international sources, analyzing content with advanced AI models including OpenAI GPT and Perplexity, and presenting critical intelligence through an intuitive interface. Our War Agent system provides autonomous threat analysis, 48-hour predictions, and multi-agent simulations to help you stay informed about escalating tensions and potential threats to global stability.
War Agent Intelligence System
🤖 Autonomous AI Orchestration
Our War Agent is a sophisticated LangGraph-based orchestration system that autonomously monitors, analyzes, and predicts global conflicts:
- • Perplexity API Integration: Real-time intelligence gathering from current news
- • OpenAI GPT Analysis: Deep analysis using OpenAI's GPT for threat assessment
- • Historical Pattern Analyzer: Analyzes 90 days of ACLED/GDELT data to identify escalation patterns
- • Multi-Stage Workflow: Monitor → Analyze → Escalate → Brief pipeline for comprehensive intelligence
- • Early Warning Signals: Detects 5 signal types: escalation, pattern match, anomaly, threshold, correlation
🎯 48-Hour Predictive Analysis
Advanced prediction capabilities using historical patterns and real-time data:
- • Pattern Matching: Identifies similar historical events from ACLED/GDELT databases
- • Escalation Probability: Calculates 24h, 48h, and 1-week escalation risks
- • Actor Behavior: Predicts responses from USA, Russia, China, NATO, Iran based on historical patterns
- • Timeline Forecasting: Generates probable event sequences with confidence scores
🎮 Multi-Agent Simulation
Sophisticated simulation engine (inspired by but not using the AGI Research WarAgent):
- • Custom War Agents: Each major actor (USA, Russia, China, etc.) has unique decision profiles
- • Interaction Modeling: Simulates diplomatic, economic, and military responses
- • Escalation Dynamics: Models how actions trigger reactions across multiple actors
- • Outcome Probabilities: Generates multiple scenarios with likelihood assessments
💬 Interactive Chat Interface
Real-time War Agent chat with persistent history:
- • WebSocket Connection: Live bi-directional communication with War Agent
- • Context Awareness: Analyzes selected events for targeted intelligence
- • Chat History: All conversations saved and linked to user accounts
- • Smart Responses: Combines real-time data with historical analysis
Data Pipeline & Analysis
1. Data Collection Layer
Continuous monitoring from multiple verified and news sources:
- • ⭐ ACLED: Armed Conflict Location & Event Data (verified conflict database)
- • ⭐ GDELT: Global Database of Events, Language, and Tone (2.0 API)
- • ⭐ Perplexity: Real-time web intelligence (via Sonar Pro model)
- • RSS Feeds: 50+ news sources updated every 3 minutes
- • Priority System: Sources ranked 1-10 by reliability and speed
2. AI Processing Pipeline
Multi-stage AI analysis with redundancy and verification:
- • Batch Processing: Processes 5 articles simultaneously for efficiency
- • OpenAI GPT: Primary analysis using GPT-4/5 models (temperature=1 for o1 models)
- • Entity Extraction: Countries, organizations, weapons, military units, key figures
- • Duplicate Detection: Multi-layer similarity scoring (title 50%, description 20%, keywords 30%)
- • Scheduled Tasks: AI scoring every 15 minutes, rescoring daily for 7-day old articles
3. Threat Assessment Algorithm
Sophisticated scoring system combining multiple factors:
- • Critical (90-100): Nuclear threats, WMD deployment, world war declarations
- • High (70-89): Invasions, major offensives, strategic bombings
- • Elevated (50-69): Military buildups, sanctions, diplomatic breakdowns
- • Moderate (30-49): Military exercises, border tensions, arms deals
- • Multipliers: Major power involvement (+15), casualties (+10-20), nuclear keywords (+30)
4. Real-time Delivery
WebSocket-first architecture for instant updates:
- • WebSocket v2: All data flows through persistent connections
- • No Polling: Events pushed instantly as they're processed
- • Channel System: Subscribe to specific data streams (news, stats, alerts)
- • State Management: Client-side caching with React Query
Technical Architecture
Frontend Stack
- • Next.js 14 (App Router)
- • React 18 with TypeScript
- • Tailwind CSS + Framer Motion
- • React Query for state management
- • Privy for authentication
Backend Services
- • FastAPI (Python 3.11)
- • SQLAlchemy 2.0 + Alembic
- • MariaDB 10.3 + Redis
- • AsyncIO task scheduler
- • WebSocket manager v2
AI/ML Stack
- • OpenAI GPT-4/5 (primary)
- • Perplexity Sonar Pro (real-time)
- • LangChain/LangGraph orchestration
- • Custom pattern analyzers
- • Historical data from ACLED/GDELT
Infrastructure
- • Docker Compose deployment
- • Nginx reverse proxy
- • GitHub Actions CI/CD
- • Cloudflare CDN
- • Android app (Kotlin)
Smart Alert System
AI-powered alerts that learn from patterns and notify you before events escalate:
Alert Types
- • Escalation warnings (confidence-based)
- • Verified threat notifications
- • Pattern match alerts
- • Anomaly detection
Customization
- • Confidence threshold settings
- • Regional focus areas
- • Actor-specific monitoring
- • Alert frequency controls
War Room Capabilities
✅ Currently Active
- • War Agent Chat - Interactive AI assistant for conflict analysis
- • 48-Hour Predictions - ML-based escalation forecasting
- • Multi-Agent Simulations - Actor behavior modeling
- • Smart Alerts - Configurable threat notifications
- • Historical Pattern Analysis - 90-day lookback insights
- • Threat Tracker - Real-time global threat monitoring with ACLED/GDELT integration
🚧 In Development
- • War Game - Strategic token-based conflict simulation
- • War Council - Community discussion forums
- • Intel Reports - Exclusive deep-dive analysis
- • OSINT Network - Crowd-sourced intelligence
What's Next
Current focus areas as of September 2025:
✅ Recently Completed
- • LangChain/LangGraph orchestration for autonomous threat analysis
- • Perplexity API integration for real-time intelligence
- • War Agent multi-actor simulations with historical patterns
- • Enhanced ACLED/GDELT integration with AI scoring
- • Smart alerts with user preferences and confidence thresholds
🚧 In Active Development
- • War Game - Strategic conflict simulation (60% complete)
- • Enhanced mobile app with push notifications
- • Satellite imagery analysis via commercial providers
- • Multi-language support for global coverage
- • Advanced entity relationship mapping
🎯 Next Quarter Focus (Q4 2025)
- • Automated X (Twitter) alerts for critical events
- • Supply chain disruption modeling
- • Economic sanctions impact tracker
- • Community-driven intelligence verification
- • Enhanced fact-checking with multiple sources
🔮 Future Vision (2026)
- • Full autonomous threat detection and alerting
- • Predictive modeling with 7-day accuracy
- • Integration with defense/security APIs
- • Professional analyst tools and dashboards
- • Open-source intelligence (OSINT) network
Want to shape our roadmap? Join the discussion on X @WarCastApp or contribute on GitHub!
Technical Implementation Notes
About War Agent Technology
We've built a custom War Agent implementation optimized for real-time analysis. Our War Agents use LangGraph orchestration with Perplexity and OpenAI APIs for advanced simulation capabilities. Each actor (USA, Russia, China, NATO, etc.) has custom decision profiles based on historical behavior patterns.
Performance Considerations
War Agent responses may take 10-30 seconds due to multiple sequential API calls: Perplexity for real-time data (30s timeout), OpenAI for analysis, and historical pattern matching against 90 days of ACLED/GDELT data. We prioritize accuracy over speed for critical intelligence assessments.
Data Privacy
All chat histories are stored securely and linked to user accounts. Anonymous users get a temporary ID. We don't sell or share data with third parties. API calls to OpenAI and Perplexity include only the necessary context for analysis, never personal information.
Support Warcast
If you value our work and can afford it, please consider donating:
Ethereum (ETH):
0xaAC0B63a317217675985CC78b7750e6DEfEFf119
Bitcoin (BTC):
bc1qvg5uvqahyarfcnxmsuds9pvvu56kyekn3hwjf3
Made with dedication by robertschmidt.dev