Loading repository data…
Loading repository data…
ruDra2916 / repository
AgenticTrade is an AI-powered trading floor where autonomous agents simulate legendary investor strategies, analyze market data, and execute trades. Built on Python and Model Context Protocol (MCP), it integrates real-time insights, portfolio management, and push notifications for a fully automated, intelligent trading experience.
A transparent discovery signal based on current public GitHub metadata.
This score does not audit code, security, maintainers, documentation quality, or suitability. Verify the repository and its current documentation before adoption.
Traders today rely on rigid bots and static signals. AgenticTrade changes the game by launching autonomous AI agents, each inspired by legendary investors like Buffett, Soros, Dalio, and Wood. These agents research markets, generate trade ideas, and execute portfolio moves—fully automated and conversation-capable.
Built with Python, OpenAI LLMs, and the MCP (Model Context Protocol), AgenticTrade simulates a full trading desk with real-time decision-making and Pushover notifications.
AI traders reviewing market conditions
Log tracer tracks all trading decisions
AgenticTrade is designed as a fully autonomous trading simulation platform, relying on tools and context agents. Here’s how it works:
Trader Initialization
Tool Use via MCP
get_share_price, push, and research.market_server.py, push_server.py, etc.).Market Awareness
market.py fetches data using the Polygon API.trading_floor.py checks if the market is open.Logging and Tracing
tracers.py logs events using write_log.Notifications
| Filename | Purpose |
|---|---|
trading_floor.py | Runs all traders in a timed loop |
traders.py | Trader logic using LLMs and MCP tools |
market.py | Polygon-based share price fetcher |
market_server.py | MCP server to respond with share prices |
push_server.py | MCP server to send push notifications |
reset.py | Resets traders to their original strategies |
templates.py | Instruction templates per trader/agent |
mcp_params.py | Tool configurations for MCP servers |
tracers.py | Logs all trace and span activity |
Create a .env file with the following keys:
POLYGON_API_KEY=your_polygon_api_key
POLYGON_PLAN=paid
PUSHOVER_USER=your_user_key
PUSHOVER_TOKEN=your_app_token
RUN_EVERY_N_MINUTES=60
RUN_EVEN_WHEN_MARKET_IS_CLOSED=false
USE_MANY_MODELS=true
pip install -r requirements.txt
python trading_floor.py
python reset.py
| Name | Role Model | Strategy Type |
|---|---|---|
| Warren | Warren Buffett | Long-term value investing |
| George | George Soros | Macro and contrarian bets |
| Ray | Ray Dalio | Risk parity + macro hedge |
| Cathie | Cathie Wood | Crypto + innovation focus |
openai
python-dotenv
requests
pydantic
asyncio
pypdf
firebase-admin
gradio (optional)
Each trader sends a push alert after finishing trades. Example:
💬 Warren bought 50 shares of BRK.B after identifying undervaluation. Portfolio remains stable with strong fundamentals.
AgenticTrade combines LLM reasoning, market intelligence, and modular tools to simulate a real-world trading desk—autonomous, explainable, and intelligent. It’s the perfect platform to experiment with financial AI agents.
📄 Technical documentation and trading logs coming soon.