Connecting an MQL5 Expert Advisor to an LLM Through a Python Server

Summary

The article presents an architecture for passing market data from an MQL5 Expert Advisor to a language model through a Python server. The EA acts as a data collector and receives a structured response, while the server calculates technical indicators, constructs prompts, calls an LLM service, and stores the API credential. It describes WebSocket communication with a TCP fallback, indicator calculations using NumPy, and sending data for multiple instruments in a batch. This separation is intended to keep model selection and credentials outside the EA and reduce blocking work inside its trading loop.

The document includes implementation details and claims about response times, costs, and portfolio scale, but does not provide a controlled benchmark or evidence that model-generated signals are profitable. It focuses on infrastructure and prompt delivery rather than a validated trading strategy. Indicator calculations and model responses may need independent checks, and moving a key to a server does not by itself establish secure deployment. The article’s performance and security claims should be read as author assertions rather than verified results.

Key ideas

• The EA sends market data to a separate Python server, which handles indicators, prompts, and model requests.
• A WebSocket connection is described with a TCP fallback for communication with the server.
• The server keeps the LLM credential outside the EA and can be changed independently of the trading client.
• The server calculates technical indicators and can process several instruments in a batched request.
• The article explains system plumbing but does not establish that LLM signals are profitable or provide controlled benchmarks.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.