Skip to main content

MCP Server Framework for MetaTrader 5. Build HTTP and stdio servers with automatic tool registration and MT5 socket communication

Project description

MCP Servers for MT5


Overview

mcp-mt5-conection is a Python library that bridges MetaTrader 5 (MT5) with the Model Context Protocol (MCP). It provides a simple and extensible framework to create tools that communicate between Python and MT5 via socket connections. Define your tools once in Python, and expose them through either FastMCP (for Claude and MCP clients) or HTTP (for custom integrations).


Main Features

Library Capabilities

  • Simple Tool Registration: Use Python decorators to define MCP tools with minimal boilerplate
  • Dual Protocol Support: Expose tools via FastMCP (for Claude) or HTTP API simultaneously
  • Socket-Based Communication: Reliable TCP socket connection between Python and MT5 with JSON serialization
  • Extensible Architecture: Abstract base classes allow custom implementations (MCP, HTTP, or custom protocols)
  • MT5 Function Access: Call any MT5 function from Python through registered tools (trading, data, charts, etc.)

Repository Structure

McpServer/
├── Src/                    # MQL5 Backend Functions
├── mcp_mt5_conection/      # Python MCP Server Package
└── Configuration & metadata files

Requirements


Installation of repo code

cd "C:\Users\YOUR USER\AppData\Roaming\MetaQuotes\Terminal\YOUR ID\MQL5\Shared Projects"
tsndep install "https://forge.mql5.io/nique_372/McpServer.git"
  • For use tsndep command requerid tsndep pacakage (avaible in pypi).. This command automatically downloads all dependencies and installs all requirements from the repositories.
  • If any part of the system is private, then it will fail... contact me so I can give you access (if it's a product, you can buy it; if you have any questions, don't hesitate to contact me).

Quick Start (for final users)

1. Install Package

pip install mcp-mt5-conection

2. Create Configuration File

Create a JSON configuration file (e.g., mt5_mcp_config.json):

{
    "general_config": {
        "host": "localhost",
        "port": 9999,
        "mode": "fast_mcp"
    },
    "fast_mcp": {
        "name": "MT5McpServer"
    },
    "http": {
        "http_port": 8000,
        "name": "MT5 HTTP Server",
        "tools_namespace": "tools"
    }
}

3. Create Your Server Script

Create a Python file with your custom tools (based on mcp_mt5_conection/server_template.py):

from mcp_mt5_conection import CMt5McpConection, CToolRegisterMCP
import json
from typing import Dict, Any

# Load configuration
config : dict = None
with open("mt5_mcp_config.json", "r") as f:
    config = json.load(f)

# Initialize connection and server
mt5_conn = CMt5McpConection(
    config["general_config"]["host"],
    config["general_config"]["port"]
)
server = CToolRegisterMCP(config, mt5_conn)

# Register your custom tools using the decorator
@server.register_tool()
def open_trade(payload: Dict[str, Any]) -> str:
    """Open a new trade position"""
    pass

# Run the server
if __name__ == "__main__":
    server.run()

4. Part of MQL5

  • Functions include
#include <TSN\\Mcp\\Main.mqh>
//+------------------------------------------------------------------+
//|                                                                  |
//+------------------------------------------------------------------+
class CMcpFuncTradeTrade : public CMcpFunction
 {
protected:
  CTrade*            m_trade;
public:
                     CMcpFuncTradeTrade(CTrade* tr)
    :                CMcpFunction(0, false, "open_trade"),
                     m_trade(tr)
   {}
                    ~CMcpFuncTradeTrade(void) {}
  //---
  void               Run(CJsonNode& param, string& res) override final;
 };
//+------------------------------------------------------------------+
void               CMcpFuncTradeTrade::Run(CJsonNode& param, string& res)
 {
  ::ResetLastError();
  res = "";
  m_trade.SetExpertMagicNumber(ulong(param["magic"].ToInt(0)));
  const bool result = (param["type"].ToString("") == "buy")
                      ? m_trade.Buy(param["lot_size"].ToDouble(0.00), param["symbol"].ToString(NULL), param["price"].ToDouble(0.000), param["sl"].ToDouble(0.000), param["tp"].ToDouble(0.0000), param["comment"].ToString(""))
                      : m_trade.Sell(param["lot_size"].ToDouble(0.00), param["symbol"].ToString(NULL), param["price"].ToDouble(0.000), param["sl"].ToDouble(0.000), param["tp"].ToDouble(0.0000), param["comment"].ToString(""));
  if(!result)
    res = StringFormat("{\"ok\":false,\"error\":\"trade_failed, last mt5 error = %d\"}", ::GetLastError());
  else
    res = StringFormat("{\"ok\":true,\"result\":%lu}", m_trade.ResultOrder());
 }
  • Main EA
#include "Functions.mqh"
//+------------------------------------------------------------------+
//| Inputs                                                           |
//+------------------------------------------------------------------+
input string InpSoketAdres = "127.0.0.1";
input uint InpSoketPort = 9999;
input int InpMsPool = 100;
input int InpMsTimeoutReadNoTls = 10000;
//+------------------------------------------------------------------+
//| Global variables                                                 |
//+------------------------------------------------------------------+
IMcpBase* g_mcp_server;
CTrade g_trade;
//+------------------------------------------------------------------+
//| Expert initialization function                                   |
//+------------------------------------------------------------------+
int OnInit()
 {
//---
#ifdef TSN_MCPSERVER_FUNC_CTS
  g_mcp_server = TSN_MCPSERVER_FUNC_CTS(THE_BOT_PLACE_USER_ID);
#else
  g_mcp_server = McpServerByLeo_Create(THE_BOT_PLACE_USER_ID); 
#endif // TSN_MCPSERVER_FUNC_CTS
//---
  ::EventSetMillisecondTimer(InpMsPool);
  g_mcp_server.AddLogFlags(LOG_ALL);
//--- Graphics / Objects
  g_mcp_server.AddItemFast(new CMcpFuncObjectCreate());
  g_mcp_server.AddItemFast(new CMcpFuncObjectCreate());
//---
  g_mcp_server.Set(InpMsPool, InpMsTimeoutReadNoTls);
  if(!g_mcp_server.Conectar(InpSoketAdres, InpSoketPort, (10 * 1000))) // 10 segundos de espera para conectarse
    return INIT_FAILED;
//---
  return(INIT_SUCCEEDED);
 }
//+------------------------------------------------------------------+
//| Expert deinitialization function                                 |
//+------------------------------------------------------------------+
void OnDeinit(const int reason)
 {
//---
  if(CheckPointer(g_mcp_server) == POINTER_DYNAMIC)
    delete g_mcp_server;
//---
 }
//+------------------------------------------------------------------+
//| Expert tick function                                             |
//+------------------------------------------------------------------+
void OnTick()
 {
//---
 }
//+------------------------------------------------------------------+
//| Timer function                                                   |
//+------------------------------------------------------------------+
void OnTimer(void)
 {
  g_mcp_server.TimerEvent();
 }
//+------------------------------------------------------------------+

5. Configure MetaTrader 5

In MT5: ToolsOptionsAllowed URLs for WebRequest

  • Add 127.0.0.1 (or your configured host)
  • Enable AutoTrading and DLL imports

6. Run Server and Connect EA

  • Start your Python server: python your_server.py --config mt5_mcp_config.json
  • In MetaTrader 5, compile and attach the McpServer EA to your chart
  • Configure EA parameters to match your host/port settings

7. Use in Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "mt5_mcp": {
      "command": "python",
      "args": ["your_server.py", "--config", "mt5_mcp_config.json"]
    }
  }
}

Now you can use Claude to interact with MT5 through natural language.


License

Read Full License

By downloading or using this repository, you accept the license terms.


Documentation


Contact


Copyright © 2026 Niquel Mendoza (nique_372).
TSN Trading Systems ecosystem.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

mcp_mt5_conection-1.0.1.tar.gz (9.9 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

mcp_mt5_conection-1.0.1-py3-none-any.whl (10.6 kB view details)

Uploaded Python 3

File details

Details for the file mcp_mt5_conection-1.0.1.tar.gz.

File metadata

  • Download URL: mcp_mt5_conection-1.0.1.tar.gz
  • Upload date:
  • Size: 9.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.10.9

File hashes

Hashes for mcp_mt5_conection-1.0.1.tar.gz
Algorithm Hash digest
SHA256 0ac43eac30ea774a09b49b04e5e70e15549b9a9c3aaa5e50d7cd8f92fbf1abbc
MD5 97496fdaadc699b6277ad336fa750b3b
BLAKE2b-256 1066847270c780311bfa8061089bbcc465567d091b4340811bf7bdc85df9005e

See more details on using hashes here.

File details

Details for the file mcp_mt5_conection-1.0.1-py3-none-any.whl.

File metadata

File hashes

Hashes for mcp_mt5_conection-1.0.1-py3-none-any.whl
Algorithm Hash digest
SHA256 d095e144f3823033811be334ec2a0130641c7ab137883529e18de354ff30de68
MD5 93e3e2018d6aae5fc7527d07d8d1aee5
BLAKE2b-256 f7466e3344f7447a6abd79dfe35756aa3f130ee467c5cfd6e46adcb2d0d2a57a

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page