Build a fully autonomous AI trading bot using TradingAgents.jl and a Forex

📌 Introduction

Ever thought about letting AI analyze the market, generate signals, and place trades while you sleep?
TradingAgents.jl is a powerful reinforcement learning framework built with Julia, helping you develop smart trading strategies. Meanwhile, a Forex VPS provides a stable, 24/7 runtime environment.

This tutorial will walk you through:

• Deploying TradingAgents.jl on a VPS
• Passing trading signals to MT5
• Executing real trades automatically

With just a one-time setup, you can build a fully autonomous trading bot that never goes offline!

🧠 Step 1: What You’ll Need

• A Forex VPS (Linux preferred; Windows optional, e.g., LightNode, Vultr)
• Julia installed (v1.6 or higher)
• Python installed (v3.8 or higher recommended)
• MetaTrader 5 (MT5) installed + a real or demo account
• Basic knowledge of Linux terminal

🔧 Step 2: Overall Architecture

We will split the system into two parts:

TradingAgents (Julia) — Generates trading signals

↓

Writes to a shared file signal.txt

↓

Python script reads signals → Executes MT5 trade via API

This bridges AI strategy output to real-world trade execution.

🟩 Step 3: Deploy TradingAgents.jl on VPS

1. Install Julia and required packages

Run the following commands on your Ubuntu VPS:

# Install Julia
sudo apt update
sudo apt install julia -y

# Launch Julia and install packages
julia
] add TradingAgents ReinforcementLearning Dates

Exit Julia and create the project directory:

mkdir ~/trading_bot && cd ~/trading_bot

2. Create strategy script agent.jl

# agent.jl
using Dates

function run_strategy()
    signal = rand(["BUY", "SELL", "HOLD"])  # Simulated signal
    price = round(rand(1.080:0.0001:1.090), digits=5)

    open("signal.txt", "w") do f
        write(f, "$(Dates.now()),$signal,$price")
    end
end

while true
    run_strategy()
    sleep(300)  # Run every 5 minutes
end

🟨 Step 4: Create MT5 Auto-Trading Script (Python)

1. Install MT5 Python Module

pip install MetaTrader5

⚠️ Note: MetaTrader 5 must run on Windows. If your VPS runs Linux, you can use Wine + GUI, or choose a dual-OS VPS (e.g., LightNode).

2. Create Python script mt5_bridge.py

import MetaTrader5 as mt5
import time
import os

MT5_SYMBOL = "EURUSD"
LOT = 0.1

# Initialize MT5
mt5.initialize()
mt5.login(account_number, password="your_password", server="your_server")

def send_order(signal, price):
    if signal == "BUY":
        order_type = mt5.ORDER_TYPE_BUY
    elif signal == "SELL":
        order_type = mt5.ORDER_TYPE_SELL
    else:
        return

    request = {
        "action": mt5.TRADE_ACTION_DEAL,
        "symbol": MT5_SYMBOL,
        "volume": LOT,
        "type": order_type,
        "price": price,
        "deviation": 10,
        "magic": 100234,
        "comment": "TradingAgentsBot",
        "type_time": mt5.ORDER_TIME_GTC,
        "type_filling": mt5.ORDER_FILLING_IOC,
    }

    result = mt5.order_send(request)
    print("Order result:", result)

# Monitor signal.txt and place order
last_signal = ""
while True:
    if os.path.exists("signal.txt"):
        with open("signal.txt", "r") as f:
            content = f.read().strip()
            if content and content != last_signal:
                timestamp, signal, price = content.split(",")
                send_order(signal, float(price))
                last_signal = content
    time.sleep(10)

🛠️ Step 5: Automate Running (Recommended: tmux)

# Start the strategy script
tmux new -s agent
julia agent.jl

# Open new tmux window to run MT5 bridge
tmux new -s mt5
python3 mt5_bridge.py

Alternatively, use supervisor or systemd for startup automation.

🌍 Step 6: Recommended VPS Providers (for this setup)

LightNode: www.lightnode.com

• Hourly billing supported
• 40+ global locations (low latency from São Paulo / Tokyo)
• Linux + Windows available
• Fast deployment, perfect for EA hosting

📦 Step 7: Advanced Suggestions

• Replace signal.txt with SQLite or Redis for better stability
• Log every signal and trade for tracking and debugging
• Use Docker to containerize and migrate the setup easily
• Add a basic dashboard UI to monitor trading status (optional)

✅ Conclusion: AI-Powered Trading Is No Longer a Dream

With this guide, you can deploy TradingAgents-based strategies on a VPS and execute real trades via MT5.
24/7 uptime, low latency, full automation, and intelligent decision-making—this is how you gain the edge over traditional traders!