Rust + ZMQ + MT5's MQL5: Exploiting Micro-Second BID/ASK Live Trading Data

The system operates on a publisher-subscriber model using ZeroMQ (ZMQ) to bridge MetaTrader 5 (via MQL5) and the external Rust application. To exploit micro-second BID/ASK live trading data to create a simple orderflow charting tool.

🌐 Software Demonstration

2026-01-21.10-45-20.mp4

Mechanics

1. Publisher (MQL5):

   • The ZmqPublisher.mq5 Expert Advisor initializes a ZMQ PUB socket and binds it to tcp://*:5555.
   • On every market tick (OnTick()), it retrieves the current Bid/Ask prices.
   • It constructs a JSON object (e.g., {"symbol": "XAUUSD", "bid": 2025.50, ...}) and publishes it as a message.

2. Subscriber (Rust):

   • The mt5-chart application initializes a ZMQ SUB socket and connects to tcp://127.0.0.1:5555.
   • It runs an asynchronous Tokio task that listens for incoming ZMQ messages.
   • Upon receiving a message, it deserializes the JSON data into a Rust struct (TickData).
   • The data is sent via an internal channel (mpsc) to the GUI thread.
   • The eframe/egui interface updates the chart and labels in real-time.

Workflow Algorithm

sequenceDiagram
    participant MT5 as MetaTrader 5 (MQL5)
    participant ZMQ as ZeroMQ (PUB/SUB)
    participant Rust as Rust Client (mt5-chart)
    participant GUI as GUI (egui)

    Note over MT5, Rust: Initialization
    Rust->>ZMQ: Connect (SUB) to tcp://127.0.0.1:5555
    MT5->>ZMQ: Bind (PUB) to tcp://0.0.0.0:5555
    
    Note over MT5, GUI: Real-time Tick Loop
    loop Every Tick
        MT5->>MT5: Get SymbolInfoTick
        MT5->>ZMQ: Publish JSON {"symbol":..., "bid":...}
        ZMQ->>Rust: Receive Message
        Rust->>Rust: Parse JSON to TickData
        Rust->>GUI: Send Data via Channel
        GUI->>GUI: Update Plot & Request Repaint
    end

Loading

---

🏆 Trivia

The very First: Exploiting micro-second BID/ASK live trading data from MetaTrader 5 (MT5) using the specific combination of Rust + ZeroMQ (ZMQ) + MetaQuotes Language (MQL5)..

I was trying to make a trading orderflows platform, and I came up with an idea: why not use MT5 as datafeed for it? Unfortunately, there are no existing ideal projects I could use as a benchmark, hence I created my own.

I've worked on and made ~100+ Python trading projects, and I always use MT5 as datafeed, but I want faster; that is why I came up with using Rust as a solution, and ZMQ binding for MT5 is a common practice. In worse case sscenario, there are no Rust + ZMQ + MT5 combination projects existed in the world, which is pretty nuts. I didn't know it only existed in my head, Lmao. So I created it with the help of Google Antigravity.

This simple orderflow tool will serve as a benchmark for future projects.

🏆 A survey for AI leading platforms

Extensive research across GitHub, Hugging Face, and consultation with 9 leading AI models confirmed that no prior public projects utilizing this exact high-performance stack existed before this implementation.

AI Survey Results

"no public projects combination of Rust, ZeroMQ (ZMQ), and MetaTrader 5 (MT5)."

Proof

2026-01-20.20-32-35.mp4	2026-01-21.14-03-26.mp4
2026-01-22.08-42-06.1.mp4	2026-01-22.10-09-23.mp4

Additional Validation:

I used leading AI platforms (Grok, Gemini, Claude, ChatGPT, MSCopilot, Perplexity, Baidu (Ernie Bot), DeepSeek, Qwen) to validate my claim for "The very First: Exploiting micro-second BID/ASK live trading data from MetaTrader 5 (MT5) using the specific combination of Rust + ZeroMQ (ZMQ) + MetaQuotes Language (MQL5)."

In addition, I ran nine different search combinations on each of four scholarly platforms: Arxiv, Google Scholar, SSRN, and Baidu Scholar. The detailed results are documented in scholar.md.

Visit my previous app to explore 284 AI platforms in one place, AiQG

During this process, some AI models produced hallucinated results. They suggested projects that only used partial pairs, such as Python with ZMQ or C++ with ZMQ, but never the full combination of all three components that I specified, namely Rust + ZMQ + MT5.

To correct this, I provided an explicit follow-up instruction stating "strictly those 3 combination". I then repeated the same query: "Show me GitHub or Hugging projects or any link that has 'Rust + ZMQ + MT5 combination' (strictly those 3), never miss a single link".

MetaTrader 5 uses MQL5 as its native language, which is built on C++. My objective was to extract live trading data from MT5 via ZMQ and process and visualize it in Rust to enable fast development of a high performance charting tool. After this clarification and repeated querying, all platforms returned the same conclusion.

"no public projects combination of Rust, ZeroMQ (ZMQ), and MetaTrader 5 (MT5)."

Since no one did that combination, at least at the time of developing/writing (Jan 20-21, 2026), I've decided to create one.

---

📂 Project Structure

These are the only necessary files that you need to have.

Rust-ZMQ-MT5
├── MQL5                        # Relocate this (individual files) into MT5's literal path directory, located outside this repository
│   ├── Experts
│   │   └── ZmqPublisher.mq5    # The "Server": Publishes Tick Data via TCP
│   ├── Include
│   │   └── Zmq
│   │       └── Zmq.mqh         # ZMQ Wrapper Library for MQL5 (64-bit compatible)
│   └── Libraries
│       ├── libzmq.dll          # 64-bit ZeroMQ DLL (Required)
│       └── libsodium.dll       # Dependency for libzmq (Required)
├── mt5-chart                   # The "Client": Rust Desktop App
│   ├── src
│   │   └── main.rs             # Main Application Logic (Subscribes & Plots)
│   └── Cargo.toml              # Project Dependencies
└── README.md                   

---

🚀 Functionality

The system consists of two main components acting in a Publisher-Subscriber model data flow:

1. The Publisher (MetaTrader 5 / MQL5)

• Role: The Data Source.
• Implementation: An Expert Advisor (ZmqPublisher.mq5) utilizing a custom ZMQ wrapper (Zmq.mqh).
• Protocol: Binds to a TCP HOST on port 5555.
• Operation: On every incoming tick (OnTick data event), it:
  • Extracts Symbol, Bid price, Ask price, and Server Time.
  • Serializes this data into a JSON string.
  • Publishes it instantly to the ZMQ socket.

2. The Subscriber (Rust)

• Role: The Data Consumer & Visualization.
• Implementation: A high-performance GUI application built with eframe (egui).
• Protocol: Connects to the TCP socket on localhost:5555.
• Operation:
  • Asynchronous Data Ingestion: A background tokio task listens for incoming ZMQ messages.
  • Deserialization: Parses the raw JSON bytes into Rust TickData structs.
  • Real-time Plotting: Updates an egui_plot chart at 60fps (or higher) to visualize the Bid/Ask spread in real-time.

---

🔧 Installation & Usage

Prerequisites

• MetaTrader 5 Client link and it must already be connected to a Broker (use whatever broker you like)
• Rust (via rustup) link and configure its file path Link on your computer.
• Visual Studio Build Tools (C++) (Required for Rust Linker) link select the C++ Build tools (in my case I've downloaded 6gb)

Step 1: MT5 Setup

1. Copy MQL5/Experts/ZmqPublisher.mq5 to your MT5 Data Folder's MQL5/Experts/. (relocate outside this repository, into your MT5 literal path directory)
   
2. Copy MQL5/Include/Zmq/Zmq.mqh to MQL5/Include/Zmq/. (do the same for this)
   
3. Crucial: Download libzmq.dll and libsodium.dll (64-bit versions) and place them in MQL5/Libraries.
   
4. Open the MT5 platform, look at the top of the screen. Go to Tools -> Options -> Expert Advisors and check "Allow DLL imports".
   
5. Open the MetaEditor5. Go to Navigator (Ctrl+d) -> Experts -> and find ZmqPublisher.mq5. Select and compile 'ZmqPublisher.mq5' once.
   
6. Go back to the MT5 platform. Drag and drop ZmqPublisher onto any chart.
   
7. Check. Go to Toolbox (ctrl+t) -> Journal section, check if it says Success Message: ZmqPublisher bound to tcp://0.0.0.0:5555

Step 2: Rust Client Setup

1. Open a terminal (for instance: Windows PowerShell, not a code editor like VS Code) in the mt5-chart folder. Or simply copy the file path

   cd mt5-chart
   
   # In my case, I will bash
   cd C:\Users\User\Desktop\VSCode\Rust-ZMQ-MT5\mt5-chart

2. Then run the application using another bash:

   cargo run --release

---

⚠️ Common Errors & Troubleshooting

This project combines three distinct technologies (Rust, MQL5, ZeroMQ) across two different runtime environments (MetaTrader 5 and Windows Desktop). This complexity can lead to specific errors. Below is a comprehensive guide to resolving every known issue.

1. Critical: Antivirus & Terminal Permissions

Symptom: Access Denied, Terminated, or the Rust window appears and immediately closes.

• Cause: Many Antivirus software (including Windows Defender) and built-in IDE terminals (like VS Code's integrated terminal) restrict programs from opening high-speed network sockets or interacting with other processes.
• Fix:
  1. Do NOT use VS Code's Intergrated Terminal. It often lacks the necessary permissions.
  2. Open a standalone Windows PowerShell instance as Administrator.
  3. Navigate to your project directory manually (cd C:\Users\...\mt5-chart).
  4. Run cargo run --release from there.

2. MQL5: "Access Violation" / Crash on Load

Symptom: MetaTrader 5 crashes instantly when you drag the EA onto a chart, or the EA simply disappears from the chart "smiling face" icon.

• Cause: 64-bit Pointer Truncation. The original MQL5 language was often used on 32-bit systems where an int (4 bytes) was the same size as a pointer. On modern 64-bit MT5, pointers are 8 bytes (long). If you cast a pointer to an int, you lose half the memory address, causing an immediate crash when the program tries to access that invalid location.
• Fix:
  • Never use int for handles. Use long or intptr_t.
  • Our provided Zmq.mqh has been patched to use long for all socket context and socket handles. Ensure you are completely replacing your old Zmq.mqh with the one from this repository.

3. MQL5: "Error 126" (Cannot Load Library)

Symptom: The Expert Advisor log shows Cannot load 'libzmq.dll' [126].

• Cause: This error means "Module Not Found". It happens for one of three reasons:
  1. Missing DLL: You didn't put libzmq.dll in the MQL5/Libraries folder.
  2. Missing Dependency: libzmq.dll itself depends on libsodium.dll and the Visual C++ Redistributable. If those are missing, libzmq.dll will fail to load, triggering the same header error.
  3. Architecture Mismatch: You downloaded a 32-bit DLL for a 64-bit MT5 terminal (or vice versa).
• Fix:
  • Ensure both libzmq.dll AND libsodium.dll are in MQL5/Libraries.
  • Ensure you downloaded the x64 (64-bit) versions of these DLLs.
  • Install the latest Microsoft Visual C++ Redistributable (x64).

4. MQL5: "Error 22" (Invalid Argument) during Bind

Symptom: The EA loads, but the Journal says ZmqPublisher: Bind failed with error 22.

• Cause 1: Binding to Wildcard. Defaulting to tcp://*:5555 works on Linux but can fail on specific Windows network configurations or Firewall setups.
  • Fix: Change the bind address in your code (or input settings) to tcp://0.0.0.0:5555 to explicitly bind to all interfaces, or tcp://127.0.0.1:5555 for local-only.
• Cause 2: String Encoding. MQL5 uses wchar_t (Unicode) for strings internally. ZeroMQ demands C-style standard ASCII/UTF-8 strings. Passing a raw MQL string to the DLL sends garbage bytes.
  • Fix: The Zmq.mqh wrapper must use StringToCharArray to convert the string to a byte array before passing it to zmq_bind.

5. MQL5: "DLL Imports Not Allowed"

Symptom: Expert 'ZmqPublisher' cannot be loaded because DLL imports are not allowed.

• Cause: Security setting in MT5.
• Fix: Go to Tools -> Options -> Expert Advisors and check the box "Allow DLL imports". This is mandatory.

6. Rust: "Linker 'link.exe' not found"

Symptom: cargo run fails with a massive wall of red text mentioning link.exe or msvc.

• Cause: Rust on Windows (MSVC version) relies on the Microsoft C++ Linker, which is not installed by default with Windows.
• Fix:
  1. Download the Visual Studio Build Tools.
  2. In the installer, select "Desktop development with C++".
  3. Ensure the "MSVC ... -C++ x64/x86 build tools" component is checked on the right side.
  4. Install (approx. 5-6GB).

7. Rust: "OS Error 32" (File used by another process)

Symptom: error: linking with link.exe failed or Access is denied when trying to run/build.

• Cause: The compiled executable (mt5-chart.exe) is currently running, or is locked by:
  • A previous instance that didn't close properly.
  • The VS Code underlying process.
  • Your Antivirus scanning the new .exe.
• Fix:
  1. Close the GUI window fully.
  2. If it persists, open Task Manager and kill mt5-chart.exe.
  3. Run cargo clean to wipe the locked artifact, then cargo run --release again.

8. Generic: ZeroMQ Message Format Mismatches

Symptom: Rust client connects but panics with Error parsing JSON or prints weird symbols.

• Cause: The MQL5 Publisher and Rust Subscriber must agree exactly on the data format. If MQL5 sends {"bid": ...} but Rust expects {"Bid": ...} (case sensitivity) or if MQL5 sends a trailing null byte \0 that JSON parsers hate.
• Fix:
  • Ensure the JSON structure key names match exactly.
  • In Rust, strip the trailing null character if necessary: msg.trim_matches(char::from(0)).

9. Generic: "Address in Use" (EADDRINUSE)

Symptom: You restart the EA and it says Bind failed.

• Cause: You cannot bind to port 5555 if something is already bound to it. Often, when you remove an EA, MT5 might not release the DLL handle instantly, keeping the socket open.
• Fix: Restart the MetaTrader 5 terminal completely to force-close all DLL handles.

---

📜 Citation

If you use this project in your research or application, please cite it as follows:

@software{RustZmqMt5,
  author = {Albeos, Rembrant},
  title = {Rust + ZMQ + MT5's MQL5: Exploiting Micro-Second BID/ASK Live Trading Data},
  year = {2026},
  url = {https://github.com/algorembrant/Rust-ZMQ-MT5},
  note = {First known implementation for exploiting micro-second live BID/ASK trading data using Rust + ZMQ + MT5's MQL5}
}