Inspiration Modern financial markets no longer trade in isolation. A delay in global shipping routes, sudden geopolitical tensions, or macro policy shifts can impact portfolio performance before traditional news outlets even draft a headline. We realized that individual traders and quants are often overwhelmed by fragmented tools—monitoring market charts in one tab, global macro indicators in another, and tracking real-world geopolitical events elsewhere.
We were inspired to build Alpha-Quant-Copilot: a unified, AI-driven quantitative copilot that bridges Open-Source Intelligence (OSINT) with dynamic market analytics, giving traders a comprehensive 360-degree situational awareness engine.
What it does Alpha-Quant-Copilot is an all-in-one AI platform designed for real-time market tracking, geopolitical awareness, and intelligent portfolio management:
OSINT Situational Radar: An interactive global map powered by Leaflet that aggregates live feeds across maritime assets (AISStream), aviation flights (OpenSky), geopolitical conflicts (GDELT), and disaster alerts (ReliefWeb/NOAA).
AI Situation Brain: Leverages DeepSeek to synthesize complex global events, macroeconomic signals, and social sentiment into actionable market intelligence in real-time.
Global Market Analytics: Offers synchronized K-line chart analysis (lightweight-charts), technical indicators, order book insights, and dynamic Take-Profit/Stop-Loss (TP/SL) engine calculation.
Smart Portfolio & Watchlist: Manages live positions, tracks PnL, performs AI-powered risk diagnosis, and triggers automated alerts over WebSocket and SSE protocols.
How we built it We built the application with a high-throughput, modern full-stack architecture:
Frontend: Next.js 15 (App Router), React 19, Tailwind CSS, Framer Motion, Leaflet, and lightweight-charts.
Backend & API Layer: Over 30 custom API routes integrating multi-source financial and macro adapters (East Money, Stooq, Finnhub, Sina, FRED, EIA, BLS, Treasury).
AI Integration: DeepSeek API with streaming response capabilities for real-time risk assessment and strategy generation.
Data & Auth: PostgreSQL managed via Supabase, ORM via Prisma 7, and authentication via Clerk.
Real-Time Data Layer: Server-Sent Events (SSE) for auto-reconnecting server pushes alongside WebSocket connections for bidirectional streaming.
Challenges we ran into Heterogeneous Data Normalization: Integrating disparate data sources—ranging from maritime AIS coordinates and OpenSky aviation logs to GDELT conflict vectors—into unified SituationalEntity models required robust backend data mapping and caching logic.
Real-time Map Rendering: Visualizing thousands of real-time geographical entities without causing UI thread freezing in Next.js required optimized Leaflet layer management and web worker decoupling.
Low-Latency Streaming: Balancing high-frequency market quote streams with deep AI reasoning models (DeepSeek) without exceeding API rate limits or overwhelming the client bundle.
Accomplishments that we're proud of Unified Multi-Domain Radar: Successfully built a fully operational OSINT dashboard that visualizes global macro risk and financial assets side-by-side.
Dynamic Engine Precision: Engine calculations that seamlessly update dynamic Stop-Loss / Take-Profit parameters based on live K-line volatility and portfolio position status.
Robust Real-time Infrastructure: Delivered zero-latency streaming setups using unified WebSocket/SSE architecture alongside Next.js 15 App Router.
What we learned How to design modular data adapter pipelines in TypeScript to seamlessly aggregate diverse third-party APIs.
Strategies for optimizing rendering performance when managing concurrent data streams across map canvases and charting libraries.
Practical prompting patterns and caching mechanisms (3-minute TTL caching strategies) to make LLMs effective for time-sensitive macroeconomic analysis.
What's next for www.alphapercept.com - Public Automated Trade Execution: Integrating broker APIs for automated order placement based on AI strategy signals.
Custom Backtesting Engine: Allowing users to backtest OSINT-driven trading strategies against historical geopolitical events.
Enhanced Multi-Agent Intelligence: Deploying specialized sub-agents dedicated to monitoring specific market sectors, commodities, and regional macro policies.
Log in or sign up for Devpost to join the conversation.