AI-Powered Multi-Agent Investment Intelligence Platform
Bloomberg Terminal meets OpenAI meets a Hedge Fund Operating System.
Retail investors and financial enthusiasts often lack access to institutional-grade tools that synthesize real-time market data, technical indicators, news sentiment, and risk analysis into an actionable format. Most available tools are either isolated screeners, simple portfolio trackers, or overly complex platforms without AI-assisted synthesis.
Stockox is an AI investment operating system that leverages an innovative AI Committee approach. Instead of relying on a single LLM prompt, Stockox coordinates multiple specialized AI agents (Research, Technical, News, and Risk) that independently analyze a stock. Their findings are aggregated by a Committee Agent to produce a final, highly confident investment recommendation.
- The AI Committee: Specialized autonomous agents perform Research, Technical Analysis, News Analysis, and Risk Evaluation to generate a unified investment decision.
- Real-Time Market Data: Integration with high-fidelity financial APIs (Finnhub, Alpha Vantage) synced via Redis cache.
- Portfolio Intelligence: Track holdings, evaluate cash balance, and monitor daily changes in an intuitive dashboard.
- Neo-Brutalist UI: A premium glassmorphism interface built with Next.js 15, TailwindCSS, and Shadcn UI, designed to look like a venture-backed fintech platform.
- Enterprise Architecture: Complete separation of concerns with a Go (Gin) backend handling all business logic and external API integrations, keeping the frontend purely presentational.
- Frontend: Next.js 15, React, TypeScript, TailwindCSS, Shadcn UI, TanStack Query, Framer Motion
- Backend: Go, Gin Framework, GORM
- Database: Supabase PostgreSQL
- Caching: Redis
- Authentication: Clerk
- External APIs: Finnhub (Future: Polygon, Alpha Vantage, OpenAI/OpenRouter)
- Deployment: Vercel (Frontend & Backend), Supabase (DB)
- Frontend strictly handles presentation and state via TanStack Query.
- Backend (Go) manages all business logic, AI agent coordination, and external API calls. 3 Caching Layer (Redis) prevents redundant API calls for quotes and news.
- Database (Supabase) is the single source of truth for portfolios, watchlists, analysis sessions, and recommendations.
(Add screenshots of your application here)
- Dashboard Overview
- Research Terminal
- AI Committee Analysis View
- Node.js (v18+)
- Go (1.21+)
- PostgreSQL / Supabase CLI
- Redis
- Clerk Account
Copy the .env.example file to your root or respective directories and populate the required keys. See .env.example for details.
- Create a Supabase project or set up a local PostgreSQL database.
- Run the SQL migration scripts located in
backend/database/schema.sql.
cd backend
go mod download
go run ./cmd/server/main.gocd frontend
npm install
npm run dev- Frontend: Deploy via Vercel connecting to your GitHub repository. Ensure UI environment variables are set.
- Backend: Deploy via Vercel (using the Go runtime configuration in
vercel.json). Set API secrets and database URLs. - Database: Hosted on Supabase.
- Cache: Hosted Redis instance (e.g., Upstash).
- Phase 1: Core Dashboard, Multi-Agent Architecture, Finnhub Integration. (Completed)
- Phase 2: Real-time WebSockets, Advanced Risk Modeling, Clerk Webhook Sync. (In Progress)
- Phase 3: Custom AI Agent Prompting, Portfolio Rebalancing Suggestions, Polygon.io Integration.
- Phase 4: Band Protocol Integration for decentralized oracle data validation.
We welcome contributions! Please review our Contributing Guide for details on our coding standards, branch strategy, and PR process.
This project is licensed under the MIT License - see the LICENSE file for details.