VibeTrading

AI-native quantitative trading research platform

---

VibeTrading is an AI-powered trading research platform that combines LLM agents with quantitative backtesting, factor research, and multi-broker connectivity. Research ideas, discover alpha, and deploy strategies — all through natural language or a real-time web dashboard.

---

✨ Features

🧠 AI Research Agent

Chat-driven market analysis, strategy development, and backtesting. Just ask "what factors drive MSFT returns?" or "backtest a mean-reversion strategy on SPY".

📊 Alpha Zoo — 460+ Factors

Pre-built library of academic alpha factors with benchmarks, correlation matrices, and performance attribution — ready to screen, combine, and deploy.

🔗 Multi-Broker Trading

Paper trade across Robinhood, Interactive Brokers, Alpaca, Binance, OKX, Tiger Brokers, and more — all from a unified interface.

⚙️ Backtesting Engine

PIT-safe fundamental data, multi-asset support, Monte Carlo simulation, and factor attribution. Walk-forward validation built in.

🐝 Swarm Intelligence

Deploy multi-agent investment committees, quant desks, and risk committees that debate, vote, and manage portfolios collaboratively.

🌐 Data Layer

18+ free market data sources with automatic failover, intelligent caching, and global coverage — stocks, crypto, FX, futures, and options.

📈 Web Dashboard

Real-time chat, run details, correlation matrices, strategy comparison, and portfolio tracking — built with React 19 and ECharts.

💬 IM Channels

Deploy agents to Telegram, Discord, Slack, WeChat, and 12+ other messaging platforms.

---

🏗 Architecture

┌─────────────────────────────────────────────────────────────┐
│                     IM Channels (Telegram, Discord, etc.)    │
└───────────────────────────┬─────────────────────────────────┘
                            │
┌───────────────────────────▼─────────────────────────────────┐
│                   Web Dashboard (React 19)                   │
└───────────────────────────┬─────────────────────────────────┘
                            │ SSE / REST
┌───────────────────────────▼─────────────────────────────────┐
│              API Server (FastAPI + LangChain)                │
│  ┌──────────┐  ┌──────────┐  ┌──────────┐  ┌────────────┐  │
│  │ Research  │  │ Strategy │  │  Risk    │  │  Portfolio │  │
│  │  Agent    │  │  Agent   │  │  Agent   │  │   Agent    │  │
│  └──────────┘  └──────────┘  └──────────┘  └────────────┘  │
└───────────────────────────┬─────────────────────────────────┘
                            │
┌───────────────────────────▼─────────────────────────────────┐
│                 Data & Execution Layer                       │
│  ┌──────────┐  ┌──────────┐  ┌──────────┐  ┌────────────┐  │
│  │  18+     │  │  Alpha   │  │Backtest  │  │  Brokers   │  │
│  │ Sources  │  │   Zoo    │  │ Engine   │  │ (8+)       │  │
│  └──────────┘  └──────────┘  └──────────┘  └────────────┘  │
└─────────────────────────────────────────────────────────────┘

---

🚀 Quick Start

pip install

pip install vibe-trading-ai
vibe-trading init
vibe-trading

Open http://localhost:8000 and start researching.

Docker

docker compose up

From Source

git clone https://github.com/DesusLove/VibeTrading.git
cd VibeTrading
pip install -e .
vibe-trading init
vibe-trading

Development Mode

For live reload during development:

vibe-trading dev  # Backend + frontend with hot reload
vibe-trading setup # Rebuild frontend assets
vibe-trading serve # Production build server

---

🧱 Stack

Layer	Technology
Backend	Python 3.11+, FastAPI, LangChain, LangGraph
Frontend	React 19, TypeScript, Vite, Tailwind CSS, ECharts
Data	pandas, NumPy, scikit-learn, DuckDB
Infrastructure	Docker, SSE streaming, MCP protocol

---

📁 Project Structure

agent/         Python backend — API server, MCP tools, CLI, skills, tools
├── src/       Core source code (agents, tools, skills, data layer)
├── cli/       Command-line interface (interactive chat, commands)
├── backtest/  Backtesting engine and utilities
├── skills/    Agent skills (research, strategy, risk, portfolio agents)
├── tests/     Unit and integration tests
└── runs/      Strategy run artifacts and logs

frontend/      React 19 dashboard — chat UI, real-time charts, strategy viewer
├── src/       React components and state management
├── public/    Static assets
└── dist/      Production build (generated by setup)

scripts/       Utility scripts for data ingestion, maintenance, deployment
tools/         Dev tooling — linting, formatting, CI helpers, code generation
wiki/          Documentation site (Markdown-based, GitHub Pages compatible)

🏗️ Architecture Overview

VibeTrading follows a modular architecture with clear separation of concerns:

1. Agent Layer (agent/src/agent/*): Specialized AI agents (Research, Strategy, Risk, Portfolio)
2. Tool Layer (agent/src/tools/*): 100+ tools for data, analysis, execution, and research
3. Skill Layer (agent/src/skills/*): Domain-specific capabilities (technical analysis, crypto, fundamentals)
4. Data Layer (agent/src/providers/*): 18+ market data sources with intelligent fallback
5. Execution Layer (agent/src/live/*): Multi-broker connectivity (Robinhood, IBKR, Alpaca, crypto exchanges)
6. API Layer (agent/src/api/*): FastAPI endpoints serving the frontend and CLI
7. Frontend (frontend/src/*): React 19 dashboard with real-time updates via Server-Sent Events

🔧 Development Workflow

Backend Development

# Start backend API server with auto-reload
vibe-trading dev --backend-only

# Run tests
pytest agent/tests/

# Lint and format
ruff check agent/
ruff format agent/

Frontend Development

# Start frontend dev server (Vite)
cd frontend && npm run dev

# Build for production
cd frontend && npm run build

Docker Development

# Development with hot reload
docker compose -f docker-compose.dev.yml up

# Production build
docker compose up --build

---

📄 License

Distributed under the MIT License. See LICENSE for more information.

---

Report Bug · Request Feature · Submit PR

_{Built with ❤️ for traders who code}