WAYANG - Autonomous Competitive Intelligence Orchestrator

Inspiration

WAYANG was inspired by how fast competitive landscapes move in the AI and SaaS industry. Companies often notice competitor moves too late, after a product launch, pricing change, hiring spike, or market expansion has already created impact. We wanted to build an autonomous system that helps teams detect those early signals sooner and turn them into actionable strategy.

Instead of creating another chatbot, we wanted to build an autonomous strategy war-room: a system that continuously monitors competitors, reasons over signals, recommends counter-moves, and sends real-time alerts when something important happens.

What it does

WAYANG is an autonomous competitive intelligence orchestrator.

It monitors competitors such as OpenAI, Anthropic, and Perplexity, extracts strategic signals, scores threat levels, generates predictions, recommends actions, and broadcasts HIGH alerts to Discord in real time.

The system uses five specialized agents:

• Crawler Agent: collects competitor signals from RSS/news, careers pages, pricing pages, and controlled demo anomaly sources.
• Signal Extractor Agent: classifies raw signals into strategic categories such as AI investment, pricing shift, talent activity, customer dissatisfaction, and regional expansion.
• Pattern Analyst Agent: correlates signals with historical memory and calculates threat scores.
• Strategy Advisor Agent: produces strategic recommendations based on the detected threat.
• Broadcast Agent: sends real-time alerts through webhook channels when the threat level becomes HIGH.

In the demo, WAYANG detects a Perplexity high-risk scenario from AI investment and pricing-shift signals, raises the threat score to HIGH, recommends a counter-pricing move, and sends a real-time Discord alert.

How we built it

We built WAYANG as a multi-agent Node.js system with an Express backend and a Next.js realtime dashboard.

The backend orchestrates the full autonomous loop:

Crawl → Extract → Analyze → Recommend → Broadcast

Built With

• axios
• cheerio
• cosine-similarity
• cursor
• discord-webhook-api
• dotenv
• express.js
• github
• javascript
• json-file-based-persistent-memory
• keyword-vector-similarity-retrieval
• next.js
• node-cron
• node.js
• react
• rss-parser
• server-sent-events-(sse)
• tailwind-css
• typescript
• zod