AI Monitor Platform is a comprehensive, real-time, interactive dashboard and global map designed to track the rapidly evolving Artificial Intelligence industry. It aggregates, translates, and visualizes AI news, regulatory updates, hardware releases, and market shifts from across the globe in one unified view.
🌐 Live Demo: aimonitor-nine.vercel.app
The platform is divided into several specialized modules to provide a 360-degree view of the AI landscape:
- 🌍 Overview (Interactive Map): This core module visualizes the geographic origin and global impact level of AI events using animated, color-coded markers. It provides an immediate bird's-eye view of where the most critical technological breakthroughs and regulatory developments are happening in real-time.
- 📰 News Feed: A clean, continuously updated data stream showing the latest AI developments across the globe. Users can dynamically filter this feed by specific industries (like Hardware or Software), severity levels, and specific timeframes to easily cut through the noise and find exactly what matters to them.
- 🧠 Models: This dedicated section tracks the rapid pace of both open-source and proprietary foundation models. It provides comprehensive details on the latest model releases, parameter sizes, architecture updates, and performance benchmarks to keep users informed on the bleeding edge of AI capabilities.
- 📈 Market: Designed to monitor broader market trends, this module tracks AI company stock performances and macroeconomic shifts driven by AI advancements. It offers a comprehensive look at how emerging AI technologies are actively disrupting and shaping the global financial landscape.
- 💹 Trading: A highly specialized interface built for quantitative analysis and automated trading benchmarks. This tab focuses on interpreting AI-driven trading signals and transforming raw market sentiment into actionable, quantitative insights to help power automated, algorithmic trading strategies.
- 📚 Sources: A transparent and robust management view displaying the diverse array of global data sources that feed the platform's intelligence. It tracks everything from major financial publications to highly specialized, localized AI research outlets from around the world.
- 🔔 Notifications: A customizable, proactive alert system designed to ensure users never miss a critical update. Users can configure this module to push immediate notifications regarding severe AI events, sudden regulatory changes, or major hardware announcements as soon as they happen.
At the core of the platform is a sophisticated, automated data engine powered by Anthropic's Claude models:
- Real-Time Automated Scraping: Continuously fetches the latest AI news from 10+ global regions (US, China, EU, Japan, Korea, etc.) directly via native-language RSS feeds.
- Claude-Powered Quant Processing: Leverages Claude (
claude-opus-4-8) to ingest raw, multi-lingual data and automatically:- Translate foreign articles into fluent English.
- Summarize lengthy news into concise, actionable insights.
- Dynamically Score the severity (Low to Critical) based on global impact.
- Quantify market impact — tag the affected public tickers, assign a directional market sentiment (bullish/bearish/neutral) and a −5…+5 impact magnitude, and distill a one-line quant takeaway per event.
- Categorize the news (e.g., Hardware, Regulation, Model releases) and pinpoint the exact geographic coordinates for the map.
- Daily Quant Market Pulse: Claude distills the last 24 hours of catalysts into a 3-bullet, markets-desk-style brief — naming the catalyst, the likely beneficiary/at-risk ticker, and the directional read.
- AI-Driven Trading Logic: The optional
live-trade-benchbackend interprets market shifts and news sentiment to power automated, benchmarked trading strategies.
The model is set via
CLAUDE_MODELin the pipeline scripts. Switch toclaude-haiku-4-5to cut cost/latency on this high-volume pipeline.
The Market and Trading tabs are powered by a self-contained quantitative analysis engine, implemented identically on both ends:
- Backend:
backend/backend/app/quant_analysis.py+ the/api/market/*and/api/trading/signalroutes. - Frontend:
frontend/src/lib/quant.ts.
The engine turns a raw price series into an interpretable signal by blending four complementary factors into a single risk-adjusted composite score (−100…+100) with a confidence level and a written analyst rationale:
| Factor | What it measures |
|---|---|
| Momentum | Multi-horizon (20-day / 60-day) trend persistence |
| Trend | Price position relative to the 50-day moving average |
| Mean-reversion | RSI(14) contrarian pressure at overbought/oversold extremes |
| Risk-adjustment | Annualized Sharpe ratio, with a volatility conviction penalty |
The score maps to a discrete signal — STRONG_BUY → BUY → HOLD → SELL → STRONG_SELL.
Always-on demo mode: if the live market backend is not reachable, the front-end transparently falls back to the quant engine running on deterministic, seeded data for a curated AI-sector universe (NVDA, MSFT, GOOGL, AMD, TSM, …). The Market Intelligence view therefore always renders real analysis instead of a connection error.
- Frontend: Next.js 15 (App Router), React 19, TypeScript
- Styling: Tailwind CSS v4, Shadcn UI
- Mapping: React Leaflet & CartoDB tiles
- State Management: Zustand
- Data Pipeline: Node.js,
rss-parser, Anthropic Claude API (@anthropic-ai/sdk) - Trading Backend (optional): Python
live-trade-benchquant engine underbackend/(run locally for the Trading tab)
git clone https://github.com/JackyJiang08/AI-Monitor-Platform.git
cd AI-Monitor/frontendnpm installCopy frontend/.env.example to frontend/.env.local and add your Anthropic API key (required for the automated news fetcher):
ANTHROPIC_API_KEY="sk-ant-your-key-here"See .env.example for the optional GNEWS_API_KEY (backfill) and NEXT_PUBLIC_API_URL (Trading tab backend) variables.
npm run devOpen http://localhost:3000 with your browser to see the result.
To run the background worker that automatically scrapes and analyzes global news every hour:
node start_live_news.jsThe Trading tab is powered by the Python live-trade-bench engine under backend/. It talks to a local API server (default http://localhost:8000, configurable via NEXT_PUBLIC_API_URL). Run that backend locally to enable the tab — it is not deployed alongside the Vercel frontend, so on the hosted demo the Trading tab is read-only/empty. See backend/README.md for how to start it.
The AI Monitor application code (the frontend/ app and data pipeline) is released under the MIT License. The bundled live-trade-bench trading benchmark under backend/ is a third-party project with its own license — see backend/LICENSE.