MetaTrader5 MCP Server

MetaTrader5 MCP Server

A Model Context Protocol (MCP) server that provides access to MetaTrader5 market data and trading functionality.

Features

• Connect to MetaTrader5 terminal
• Retrieve symbol information and market data
• Get historical rates (OHLCV) for any timeframe
• Access tick data and market depth (DOM)
• Real-time market data streaming
• Support for all major trading instruments

Installation

1. Install MetaTrader5 terminal
2. Install Python dependencies:

Configuration

Copy .env.example to .env and configure your MetaTrader5 credentials:

Edit the .env file with your broker credentials (optional - you can also connect without pre-configured credentials).

Usage

Starting the MCP Server

Available Tools

Market Data

• get_symbols(search_term, limit) - Smart search for trading symbols (CSV output)
• get_symbol_groups(search_term, limit) - Get available symbol groups (CSV output)
• get_symbol_info(symbol) - Get detailed symbol information
• get_rates(symbol, timeframe, candles, start_datetime, end_datetime, ohlcv, ti) - Get historical OHLCV data in CSV format
• get_ticks(symbol, count, start_datetime) - Get tick data in CSV format
• get_market_depth(symbol) - Get market depth (DOM)
• get_indicators() - List supported technical indicators (dynamic from pandas_ta)

Example Usage

The server automatically connects to MetaTrader5 when any tool is called. Connection credentials can be configured via environment variables or the server will use the terminal's current login.

Intelligent Symbol Search

The get_symbols function uses a smart 3-tier search strategy:

Search order:

1. Group name match (e.g., Majors, Forex)
2. Symbol name match (e.g., EUR → EURUSD, JPYEUR, ...)
3. Description match (symbol or group path)

Parameters:

• search_term: Optional search term. If omitted, lists visible symbols.
• limit: Optional maximum number of symbols to return.

Response format:

• CSV with header name,group,description and one row per symbol.

Date Inputs

Flexible date/time parsing powered by dateparser:

• Accepts natural language (e.g., yesterday 14:00, 2 days ago)
• Accepts common formats (e.g., 2025-08-29, 2025/08/29 14:30, 2025-08-29 14:30 UTC)
• Times are interpreted and converted to UTC internally

Usage patterns for get_rates:

• Only start_datetime: returns bars forward from the start (up to candles).
• Only end_datetime: returns the last candles bars before the end.
• Both start_datetime and end_datetime: returns bars within the range (ignores candles).

Quote Data Format

Historical rates and tick data are returned in compact CSV format with intelligent column filtering:

Historical Rates (get_rates)

By default returns: time,open,high,low,close, and optionally tick_volume,spread,real_volume if meaningful.

Selecting OHLCV subset:

• Use ohlcv with letters from {O,H,L,C,V} to choose columns (time is always included).
• Examples: ohlcv=OC → time,open,close; ohlcv=V → time,tick_volume

Technical indicators via ti:

• Pass a comma-separated list like: ti=sma(14),rsi(14),ema(50)
• Supported basics: sma(length), ema(length), rsi(length), macd(fast,slow,signal), stoch(k,d,smooth), bbands(length,std)
• Indicator columns are appended to the CSV after the requested OHLCV columns.

Note: spread and real_volume columns excluded because they contained only zeros

Tick Data (get_ticks)

Core columns: time,bid,ask (always included)
Optional columns: last,volume,flags (included only if they contain meaningful data)

Note: last, volume, and flags columns excluded because they contained no meaningful data

Intelligent Column Filtering

• Meaningful Data Check: Columns are included only if they have at least one non-zero value OR multiple different values
• Space Efficiency: Empty/constant columns are automatically excluded to reduce response size
• Consistency: Core OHLC columns (rates) and bid/ask columns (ticks) are always included

Response Format

Both functions return:

Timeframes

Supported timeframes for historical data (MetaTrader5):

• Minutes: M1, M2, M3, M4, M5, M6, M10, M12, M15, M20, M30
• Hours: H1, H2, H3, H4, H6, H8, H12
• Days/Weeks/Months: D1, W1, MN1

Requirements

• Windows OS (MetaTrader5 requirement)
• MetaTrader5 terminal installed
• Python 3.8+
• Active internet connection for real-time data

Error Handling

All functions return structured responses with success/error status:

Security Notes

• Store credentials securely in environment variables
• Use demo accounts for testing
• Never commit credentials to version control
• Consider using read-only API access where possible

License

MIT License