Article-22388-Telegram-Voice-Trading-MT5-Accessibility/README.md

# Article-22388-Telegram-Voice-Trading-MT5-Accessibility

This repository is an article-derived reference project based on the original MQL5 article. It does not claim to reproduce the full original source code unless files are explicitly attached.

## Overview

This project documents a semi-automated remote voice-command trading pipeline for MetaTrader 5. The architecture links a Telegram voice message to a local Python middleware, which transcribes and parses the message, exposes commands over HTTP, and lets an MQL5 Expert Advisor poll, execute, and return results.

The design targets low-bandwidth remote control and accessibility use cases, especially when the trader is away from the desktop terminal.

## Original Article

- **Article ID:** 22388
- **Author:** Clemence Benjamin
- **Publication date:** 2026.05.26
- **Category:** Python / Experts
- **URL:** https://www.mql5.com/en/articles/22388

## Repository Purpose

This repository serves as a reference reconstruction of the solution described in the article:

- a **Python middleware** that combines:
- Telegram bot polling
- voice-note download and conversion
- speech-to-text transcription
- natural-language command parsing
- local HTTP server endpoints
- an **MQL5 Expert Advisor** that:
- polls the middleware with `WebRequest`
- parses JSON commands
- executes trades through `CTrade`
- posts execution results back to the middleware

It is useful as a study reference for:
- Telegram-to-MT5 integration
- pull-based command delivery in MQL5
- low-bandwidth remote trade control
- accessibility-oriented trading workflows

## Key Concepts

- Pull-based HTTP polling from MQL5
- Telegram voice-note handling
- Speech recognition via Python
- Audio conversion with `ffmpeg` and `pydub`
- JSON-based command exchange
- Cross-language integration between Python and MQL5
- Remote order execution with `CTrade`
- Simple in-memory queueing and result synchronization
- Risk reduction through default micro-lot sizing

## Algorithm / Architecture Summary

1. A user sends a Telegram voice message.
2. The Python bot downloads the `.ogg` file.
3. The middleware converts audio to 16 kHz mono WAV.
4. Speech is transcribed using Python speech recognition.
5. Parsed text is converted into a JSON command such as:
- `BUY`
- `SELL`
- `CLOSE_ALL`
- `BALANCE`
6. The command is placed into an in-memory queue.
7. The MQL5 EA polls `GET /get_command` every 2 seconds.
8. If a command is available, the EA:
- parses JSON using `JAson.mqh`
- executes the requested action
- builds a JSON response
9. The EA sends the result to `POST /post_response`.
10. The Telegram bot returns the outcome to the user.

Important implementation note highlighted in the article:
- `StringToCharArray` in MQL5 appends a null terminator, which must be stripped before POSTing JSON to Python.

## Mentioned or Attached Files

### Explicitly attached files

- `telegram_trading_bot.py` — Python 3 middleware script
- `OvercomingAccessibilityIV.mq5` — MQL5 Expert Advisor
- `JAson.mqh` / `Jason.mqh` — third-party JSON parser referenced from MQL5 CodeBase

### Mentioned in text

- `MQL5\Include\OvercomingAccessibilityIV\JAson.mqh`
- `MQL5\Experts\OvercomingAccessibilityIV.mq5`

## Statistics

- **Polling interval:** 2 seconds
- **Typical round-trip latency:** 1.5-4.2 seconds
- **HTTP round-trip + trade execution:** 50-200 ms
- **Bandwidth per poll request:** ~200 bytes
- **Empty queue response size:** ~20 bytes
- **Full trade cycle bandwidth:** less than 30 KB
- **Voice download size:** ~10-30 KB
- **Python version:** 3.10+
- **HTTP server bind address:** `127.0.0.1:8082`

## Tags

`MQL5` `MetaTrader5` `Telegram` `Python` `WebRequest` `Voice-Control` `Accessibility` `JSON` `CTrade` `Speech-Recognition`

## Difficulty

**Intermediate**

Requires familiarity with:
- MQL5 Expert Advisor development
- Python environment setup
- HTTP/JSON communication
- Telegram bot configuration
- Windows dependency setup for `ffmpeg`

## Limitations

- This is an article-derived reference project; the README reflects the article content and attached file list.
- Full original repository structure is not available unless the attached files are actually present in this processed input.
- The solution depends on a continuously running Windows machine with MetaTrader 5 open.
- The architecture uses polling, so execution latency includes the polling interval.
- The speech recognition path described in the article uses external Google STT through Python libraries.
- The JSON parser include is third-party and must be obtained separately if not bundled.
- If no attached source files are present in this repository snapshot, then no attached source files were available in the processed input.

## Reference

- MQL5 article: https://www.mql5.com/en/articles/22388
- MQL5 CodeBase reference for JSON parser: https://www.mql5.com/en/code/13663