signal

everything about today, in a single place.

Comment

Signal — Global Intelligence Terminal

Inspiration

Most platforms show what happened: news reports events, financial terminals show prices, and social media amplifies reactions. None clearly show what people collectively believe will happen, backed by real money and data.

Signal turns prediction markets into an accessible intelligence system by combining forecasts, macro data, and sentiment into a single live terminal.

What It Does

Signal is a full-stack intelligence terminal that unifies prediction markets, macroeconomic indicators, asset prices, and social sentiment.

Core Modules

  • Signal Feed — Live stream of AI personas (bullish, bearish, contrarian) generating market takes grounded in real data with cited sources.
  • Briefing — Daily digest summarizing sentiment shifts, key headlines, and market changes.
  • Markets — Prediction-market dashboard showing probabilities, volume context, and signal strength across crypto, economy, politics, geopolitics, sports, and tech.
  • Economy — Macro context layer linking market moves to assets and economic indicators.

Built with a terminal-style interface inspired by professional trading platforms, but accessible without enterprise costs.

How It Works

Backend

  • Python data pipeline hosted on Zerve
  • Aggregates top Polymarket markets
  • Classifies markets into six verticals
  • Enriches data with:
    • FRED macro indicators
    • Yahoo Finance asset prices
    • Google Trends attention data
  • Computes a composite Trending Score (volume, attention, momentum, signal strength)
  • Served via FastAPI endpoints

Frontend

  • Next.js 16 + TypeScript + Tailwind
  • Deployed on Vercel
  • xAI Grok API for AI persona analysis
  • X/Twitter API for real-time sentiment correlation
  • Radix UI + shadcn/ui + Framer Motion animations

Key Challenges

  • Normalizing inconsistent Polymarket pricing formats without pipeline failures
  • Handling unreliable financial APIs using caching and rate limits
  • Designing AI personas that produce distinct, non-repetitive viewpoints
  • Keeping the UI dynamic and “live” even during low market activity

Accomplishments

  • Built and deployed a complete live full-stack system solo
  • Near-real-time data enrichment pipeline with aggressive caching
  • Custom trending score that surfaces emerging signals, not just volume
  • Interactive AI-driven terminal experience designed for live demos

What I Learned

Prediction markets become significantly more useful when combined with macroeconomic and attention data.

AI-generated analysis only builds trust when it is transparently grounded in real, verifiable sources.

What’s Next

  • User accounts, watchlists, and alert webhooks
  • Multi-market aggregation (Manifold, Kalshi, Metaculus)
  • CLI interface for power users
  • Mobile-native version

Links

Live Demo: https://singnal-frontend-jqhh.vercel.app
Repo: https://github.com/404kaushik/signal-zerve-hackathon

Built With

Share this project:

Updates