EA31337 Libre: Free, Readable MT4/MT5 EA Framework

Most “free EA” downloads are a closed box.

You get a compiled binary, a screenshot of someone else’s equity curve, and a promise. You cannot read the code that decides when to buy. You cannot see how much it risks on a trade. You cannot tell whether it holds one position or quietly stacks twenty against you.

That is the problem this page is about. The free expert advisors most traders find are not tools you can inspect — they are black boxes with a sales page bolted on. And a black box inside your trading account is not a shortcut. It is a liability you cannot see.

What does a black-box EA look like when it goes wrong?

This is an illustrative composite, built from the failure pattern common to closed robots, not a report about a specific trader. A free robot runs for a few quiet days, then a strong gold (XAUUSD) move arrives. The robot keeps adding to losing positions, the drawdown deepens, and the robot stops trading without warning. Nothing in the screenshots showed any of it. Because the file was compiled, the rule that caused the damage could not be read, so nothing could be checked before the loss.

The cost that outlasts the deposit

The lost money is the obvious cost. The quieter one is that you learned nothing you can reuse.

A closed EA teaches you a single fact: whether that black box made or lost money in one particular week. It does not show you why. The entry rule, the position size, the exit, the maximum exposure — all of it stays hidden, so you cannot carry any of it into your next decision. You are exactly as good at judging the next robot as you were before you downloaded this one.

That is the real trap. Free-but-opaque software is not cheap. It moves the price from your wallet to your time, then charges you again the next time you install something you cannot inspect.

There has to be a third option between “buy a tuned robot you cannot see inside” and “write an expert advisor from scratch”. There is, and it is called an open-source framework. EA31337 Libre is one of the clearest examples on MetaTrader.

If your goal is a shortlist of finished, reviewed robots, our best MT4 EA roundup and the top ranking table are the right pages. If your goal is to understand how automated trading actually works — and to own something you can inspect and change — keep reading.

What exactly is EA31337 Libre?

EA31337 Libre is a free, open-source, multi-strategy trading robot for MetaTrader 4 and MetaTrader 5, written in MQL and released under GNU GPL-3.0. It ships with the full source code, so you can read and change every rule instead of trusting a compiled binary.

The project’s own description is short: it aims to deliver a simple implementation of the EA31337 framework and the EA31337 strategies. Those are two separate repositories. The framework (EA31337-classes) holds the shared machinery — order handling, position management, indicator wrappers. The strategy library (EA31337-strategies) holds the decision rules. Libre wires the two together into one robot you can compile and run.

Concrete numbers, because they matter more than adjectives:

• Over 35 strategies ship with the robot, and each one can analyse the market on a different timeframe independently.
• The analysis is built on standard technical indicators you already recognise — moving averages, RSI, Bollinger Bands, MACD, Stochastic and others.
• The code is compatible with both MQL4 and MQL5, so the same project runs on either terminal.
• You are free to write your own custom strategies and add them to the set.
• The repository is public on GitHub, so you can copy it as a template and check its current activity yourself.

What you are holding, then, is not a single expert advisor. It is a framework with a stack of strategies wired in. That word — framework — is the whole point of this page, and it is the difference between a tool you own and a tool you rent.

A framework is a starting point you can inspect, not a finished product you can trust blindly.

What does a framework give you that a tuned commercial product does not?

Answer first: a framework gives you three things a tuned commercial EA cannot — readable source code, strategies you can switch on or off, and a risk model you can audit. A commercial product gives you its conclusions. A framework gives you its reasoning.

1. Source code you can actually read

With a compiled EA, the strategy is a claim. With EA31337 Libre, it is code you can open in MetaEditor and step through.

You can find the exact condition that opens a trade. You can see how position size is calculated. You can check whether the stop-loss is fixed, ATR-based, or missing entirely. None of that is hidden, so none of it has to be taken on faith. When a vendor tells you their EA “manages risk intelligently”, you normally have to guess what that means. Here, you can look.

That auditability is the most underrated feature in free software. It does not make a strategy profitable. It makes a strategy knowable, which is the only honest starting point.

2. Strategies you can enable, disable or replace

A finished commercial EA is one opinion, frozen. EA31337 Libre hands you a stack of strategies and lets you decide which ones run.

Because each strategy can work off its own timeframe, you can run a fast, short-term rule on M5 while a slower rule reads H4. You can switch off the ones you have not tested. You can replace one entirely with your own logic without rewriting the order-management layer — the part most people get wrong when they build an EA by hand.

This is what “multi-strategy” buys you in practice: the ability to build a book of systems instead of betting everything on a single rule. It is also work. Bolting together strategies that all lose money in the same market conditions gives you a bigger version of the same weakness, not diversification.

3. A risk model you can audit

Most retail blow-ups in automated trading come from three things: position sizing that grows too fast, systems that add to losing trades (grid or martingale logic), and no hard cap on how many positions can be open at once.

In a closed EA, you discover all three after the fact, usually with a screenshot of the damage. In a readable framework, you can search the source for them before you attach it to anything. You can see whether the code averages down, whether it multiplies lot size, and whether it limits exposure. You can change any of it and recompile.

That is what an auditable risk model means in plain terms. Not a safer system by default — a system whose risks you are allowed to see and correct.

The trade-off, stated plainly

Option	What you get	What it costs you
Tuned commercial EA	A developed edge, hidden logic, a price tag	Visibility — you cannot inspect the rules
Open-source framework	Full visibility, control, no promises	Work — you do the testing and tuning
Build from scratch	Total control	Months of effort, most people never finish

EA31337 Libre sits in the middle column, and it is honest about being there.

How does EA31337 Libre split strategy from execution?

Answer first: it separates the strategy — the rule that decides to trade — from the execution layer — the code that places and manages orders. That split is what makes the project a framework rather than a single robot.

In a typical home-made EA, the two are tangled together. The entry condition, the position size, the stop-loss and the exit all live in one file, often in one function. Change the entry rule and you risk breaking the order handling. That is why so many DIY robots are never finished: every improvement threatens the machinery around it.

EA31337 Libre organises things differently. The framework repository (EA31337-classes) owns the shared parts — talking to the terminal, opening and closing positions, tracking the order pool, wrapping indicators. The strategies repository (EA31337-strategies) owns the decisions — when to signal a buy, a sell, or nothing. Libre itself is the thin layer that wires the two into a buildable expert advisor.

The practical result is that a strategy becomes a small, self-contained unit. You can add one, remove one, or swap one out without touching the code that manages money and orders. For someone learning automation, that is the most useful structural lesson in the project, because it is the shape professional trading systems are built in.

There is a second benefit hidden in the same design. Because the framework abstracts the terminal, the same strategy source can be compiled for MetaTrader 4 and MetaTrader 5. MQL4 and MQL5 are not interchangeable, so most developers maintain two products or support only one. Writing once and building twice is a genuine engineering saving, not a marketing line.

None of this makes the strategies good. It makes them modular, which means you can judge and replace them one at a time instead of treating the whole robot as an indivisible bet. That is the difference between a framework you can improve and a product you can only accept or discard.

What does EA31337 Libre not give you?

Answer first: it does not give you a tuned edge, ready-made presets, or commercial support. The author says so directly — the project is intended for education and research, and is not suitable for real trading without relevant knowledge.

The project’s README states that it is for education and research and is not suitable for real trading without relevant knowledge.

Read that as a feature, not a flaw. A vendor who told you the truth this plainly would sell fewer copies. Here is what each part of it costs you in practice.

• No tuned edge. The strategies are raw building blocks. Nothing is optimised for your broker, your symbol, or your timeframe. The edge, if there is one, is whatever you create by testing and combining.
• No presets. There is a sets/optimize folder for optimiser inputs, but you generate your own set files. There is no set-and-forget configuration, and no “install and profit” story.
• No commercial support. Help comes from GitHub discussions, issues, and a Telegram channel. That community is genuinely useful, but there is no service-level promise and no guarantee anyone answers in your timezone.
• No plug-and-play for live accounts. The project recommends MetaTrader 5 for backtesting, and backtesting cannot reliably simulate future outcomes.
• No warranty. Like all GPL software, it is provided as-is. If it loses money, the licence and the author say the same thing: the outcome is yours.

What it gives you	What it does not give you
Full, readable MQL source	A strategy tuned for your account
35+ strategies you control	Optimised preset files
MetaTrader 4 and 5 support	A promise it makes money
A community on GitHub and Telegram	Commercial support with a service-level promise
GPL-3.0 freedom to modify and share	Any warranty at all

Why does GPL-3.0 matter if you modify and redistribute it?

Answer first: GPL-3.0 is a copyleft licence. You can use, change and share the software for any purpose. The condition is that any version you distribute must stay free under the same licence.

Two words carry all the weight: “distribute” and “copyleft”.

Distribute is the trigger. If you change the code and keep it to yourself — on your demo account, on your own live account — you have no obligation to publish anything. Modify it freely. The licence simply does not reach private use.

Copyleft is what happens when you hand it to someone else. The moment you share or sell a modified version, GPL-3.0 requires you to:

1. Licence that version under GPL-3.0 as well.
2. Keep the original copyright and licence notices in place.
3. Make the corresponding source code available to whoever receives it.
4. Add no extra restriction that removes a freedom the licence already grants.

That is why you will not find a GPL expert advisor sold as a closed premium product with the logic hidden. The licence forbids closing it back up. It protects the commons: whatever anyone builds on the project publicly stays open for the next person.

What the licence does not require is just as important. It does not force you to publish your private changes, it does not stop you trading your own modified build on your own account, and it does not demand royalties or a fee. GPL-3.0 is not a paywall or a permission slip you have to request. It is a condition that travels with the code: share the freedoms you were given when you pass the work on.

How GPL-3.0 compares to the permissive licences

Not every open EA project works this way, and the difference changes what you can legally do with a fork.

Licence	Copyleft?	Must you share changes you distribute?	Commercial use
GPL-3.0 (EA31337 Libre)	Yes	Yes	Allowed
MIT (many MT5 libraries)	No	No	Allowed
Apache-2.0 (many small EAs)	No	No	Allowed

The practical reading:

• If you want to sell a modified EA as a closed product, GPL-3.0 is the wrong licence for you. Choose a permissive one, or open your fork.
• If you want a permanent, inspectable tool you can keep, copyleft is a feature, because nobody can take the community’s work, improve it, and close it off.

If you want to compare licence terms across free downloads, read each repository’s licence file before you use it, and see how the TeknoTrader MQL4 indicators page handles its Apache-2.0 terms. Always check the licence on the specific project, not the category it sits in.

One more clause matters to anyone trading money: GPL-3.0 disclaims warranty. The software comes as-is, and no one is liable for what it does on your account. Read that as the honest baseline, not a technicality.

How do you evaluate EA31337 Libre on demo before risking real money?

Answer first: put it in the MetaTrader 5 Strategy Tester on your own broker’s data, then forward-test it on a demo account for several weeks, changing one variable at a time. No framework earns trust by reading alone.

Here is the sequence that actually filters out bad configurations instead of flattering them.

1. Start in the Strategy Tester, not on a live account. The author recommends MetaTrader 5 because its tick modelling is more accurate than MetaTrader 4’s. Same code, better evidence.

2. Understand what modelling quality does and does not tell you. A 99% modelling-quality backtest means the price data was detailed. It does not mean the strategy is valid. High quality with a bad idea is still a bad idea.

3. Backtest on the broker you will actually trade. Spread, commission, swap and execution differ between brokers. A result on one feed is not the same system on another’s.

4. Change one parameter at a time. If an improvement only appears at one specific setting, you have fitted noise, not found an edge. Real edges survive small changes; curve-fits do not.

5. Hold back data you never optimised on. Tune on one period, then check the result on a later period you did not touch. If it collapses out of sample, it was fitted.

6. Forward-test on demo for at least a month. You want quiet weeks, volatile weeks, and a news event or two. Demo fills are not real fills, but a month of forward behaviour still exposes logic that backtests hide.

7. Record the numbers, not the vibes. Track drawdown and profit factor with a tool such as Myfxbook so you have a written record. Decide your maximum acceptable drawdown before you start, and stop if it is reached.

Evaluation check	What you are looking for
Same code, both terminals	A clean compile on MetaTrader 4 and 5
Out-of-sample test	Performance that survives unseen data
Broker-specific data	Results that hold on your real spread
One-variable changes	Robustness, not a single lucky setting
Forward demo month	Behaviour across different market conditions
Position limits	A visible cap on open trades

The point of all this is not to find a robot that always wins — that does not exist. It is to find out, cheaply and on demo, whether a system behaves the way you expect when the market stops cooperating. That is the question a compiled EA lets you answer only with real money.

Which mistakes do traders make with a free framework?

Answer first: they skip the backtest, trust the code because it is open, and treat “free” as “safe”. Openness removes one risk — opacity — and leaves every other risk exactly where it was.

The most common errors, roughly in order of how much they cost:

• Assuming open source means safe. Readable code is not good code, and good code is not a profitable strategy. You still have to test the logic, because a clean, honest rule can still lose money steadily.
• Running it live before forward-testing. Backtest results, however detailed, are not live behaviour. A month on demo costs nothing and catches the things a backtest hides.
• Ignoring the grid and martingale question. Search the source for averaging-down and lot multiplication before you accept any equity curve. A smooth curve built on stacked losing positions is a drawdown waiting for the right trend.
• Over-optimising. The more parameters you tune to one history, the worse the system tends to do on the next one. Fewer knobs, set robustly, usually beats a perfect fit.
• Betting on one strategy. The framework’s value is combination — several rules that fail at different times. Running a single strategy you found in the repo throws that away.
• Forgetting the licence. If you plan to redistribute what you build, GPL-3.0 has terms. Know them before you ship a fork.

Every one of these is avoidable with demo work and a little reading. None of them is fixed by choosing a different download.

Who is EA31337 Libre for — and who should skip it?

It is for people who want to build understanding, not buy it. Specifically:

• MQL developers who want a working multi-strategy framework to study or extend.
• Traders learning automation who would rather read real source than trust a black-box binary.
• Researchers combining strategies and wanting a free, licence-clear base to test against.

It is not for you if:

• You want a finished EA to attach to a funded account and leave running. The author says no, and we agree.
• You will not backtest. A framework rewards testing and punishes guessing.
• You expect presets, signals, or someone to tune it for you. None of that is in the box.

If those last three sound like you, a framework is the wrong tool, and that is fine. A curated shortlist is the right tool — our top ranking page ranks finished robots, and if you would rather mirror other traders than run your own code, copy trading is a completely different route.

What do you actually get when you download it?

You get the whole project from its origin, with nothing held back and nothing charged. There is no upsell, no licence key, and no “pro” tier where the good version lives.

Spec	Detail
Product	EA31337 Libre
Type	Multi-strategy expert advisor framework
Platforms	MetaTrader 4 and MetaTrader 5
Languages	MQL4 / MQL5 from a single codebase
Strategies	35+, individually configurable, per-strategy timeframes
Source code	Included — full MQL source
Underlying projects	EA31337 framework and EA31337 strategies
Licence	GNU GPL-3.0
Author	EA31337 project
Price	Free
Hosting	External — the source repository on GitHub
Recommended testing	MetaTrader 5 Strategy Tester, then a demo account

The repository itself is organised the way a real software project is: a src folder for the code, a sets/optimize folder for optimiser inputs, a docs folder, and a .gitmodules file that pulls in the framework and strategy libraries. That structure is part of the value — you are not just getting an idea, you are getting a codebase you can build on.

Because it is hosted at the source, you always get the current version rather than a copy that has drifted. The download button below links straight to the project.

And the same honest list from earlier still stands, because it is the reason this page exists:

• It is not a tuned edge, and it is not a preset configuration.
• It is not supported as a commercial product — help is community-based.
• It is not suitable for real trading without relevant knowledge, in the author’s own words.

What it is: a free, inspectable, GPL-3.0 framework you can read, compile, test, and modify. For a lot of traders, that is worth more than another tuned robot they cannot see inside. For our broader library of tools you can actually inspect, the free download hub collects them in one place, and the blog covers the testing method end to end.

Your next move: read the source before you risk a cent

One action. Open the EA31337 Libre repository, read the strategy code, compile it in MetaEditor, and run it in the MetaTrader 5 Strategy Tester on a single symbol. Then move it to a demo account and let it run.

That is the whole play. Not because the robot will be profitable — nobody can tell you that, and anyone who does is selling something. Because by the end of it, you will understand what this expert advisor does, how it sizes positions, and how it exits, instead of guessing from a screenshot.

A framework is a starting point you can inspect, not a finished product you can trust blindly. EA31337 Libre is a strong, genuinely free starting point for anyone who wants to stop taking software on faith. Whether you keep it or replace every strategy with your own, the source is right there — and that is the part no closed EA can ever give you.

Before anything real: test on demo first. Trading forex and CFDs carries a high risk of loss and is not suitable for everyone, and leverage can work against you as easily as for you. Nothing on this page is financial advice, and neither we nor the original author accept liability for trading losses. Only ever risk money you can afford to lose.