R4CopierA trade copier reproduces the orders from one trading account on other accounts, automatically. This guide covers what that actually involves — how a copy is executed, the difference between EA-based and API-bridge copiers, and the full comparison between cloud-hosted and self-hosted architectures.
It is a small idea with a lot of consequences. The idea first.
A trade copier is software that monitors one trading account and reproduces its activity on one or more other accounts. When the monitored account opens a position, the copier opens the same position elsewhere. When it closes, modifies a stop, or takes partial profit, the copier does that too — typically within tens to hundreds of milliseconds, and without anyone watching.
The vocabulary is near-universal across products. The account being watched is the master (also: source, provider, leader). The accounts receiving orders are slaves (also: followers, receivers, targets). One master can feed many slaves; a slave can, in most products, follow only one master at a time — and where it can follow several, you have introduced a position-netting problem that is worth understanding before you enable it.
The thing a copier does not do is decide anything. It has no strategy, no opinion and no signal of its own. Every trade it places is a trade a human or an expert advisor already placed on the master. A copier is plumbing: it moves a decision that has already been made from one place to several others, faster and more reliably than retyping it.
The one-line version: a trade copier is a synchroniser for trading accounts. Everything that follows in this guide — latency, credentials, symbol mapping, lot sizing — is a detail of how faithfully and how safely it manages to keep those accounts in sync.
They have almost nothing in common except the need to place one decision in several places at once.
By far the largest group. A trader passes several evaluations and ends up managing three, five, or a dozen funded accounts across different firms — each with its own balance, drawdown ceiling and rules.
Placing every trade a dozen times by hand is not merely tedious; the fills drift apart, the sizing goes wrong under pressure, and one fat-finger breaches a daily-loss limit.
A strategy is running on a small live account and being trialled on a larger one, or on a second broker with different execution. Copying keeps both in step so the comparison is actually a comparison.
Someone trading their own account while subscribers mirror it. Here the copier is the product, and reliability under load — dozens of slaves, one fill — is the requirement that decides which product survives contact with reality.
Split across brokers deliberately: for jurisdictional reasons, for redundancy if one platform goes down mid-session, or because swap and spread differ enough to matter. A copier keeps the split from becoming a second job.
What these have in common is that the decision has already been made and the work left is purely mechanical. That is exactly the class of work worth automating — and exactly the class where a subtle failure, like a copier that silently stops after a broker disconnection, is expensive rather than merely annoying.
Five steps. Every copier on the market does these; they differ in where each step runs and how well it handles the edge cases.
The copier has to learn that something happened. Depending on the architecture, it either polls the master terminal’s open-position list several times a second, or it receives an event callback the moment the platform registers a fill.
This is the first place products diverge in quality. Polling every 500ms means your worst case is 500ms of blindness before the copy even begins. Event-driven detection is effectively instant but requires deeper integration with the platform.
The master traded EURUSD. The receiving broker may call it EURUSD.pro, EURUSDm, fxEURUSD or EURUSD_i. Gold might be XAUUSD on one and GOLD on the other. Indices are worse still.
A copier resolves this with prefix and suffix rules, plus an explicit override table for the instruments that do not follow any pattern. Getting this wrong is the single most common reason a newly configured copier places nothing at all — or, much worse, places the right size on the wrong instrument.
A copy is almost never the same volume as the original. A $200,000 funded account and a $5,000 personal account cannot take the same 2-lot position and remain within their respective risk rules. The copier applies a rule per receiving account, usually one of:
Fixed multiplier — master lot × a constant. Predictable; goes wrong as soon as balances drift apart. Percentage of balance or equity — sizes each account against its own capital, so a 1% risk on the master is 1% everywhere. The right default for most people. Static volume — always the same lot regardless of the master. Useful for rigid prop-firm rules and for testing.
Then the result gets clamped: rounded to the broker’s lot step, floored at its minimum, capped at its maximum. A copier that skips the clamping stage produces rejected orders on exactly the accounts where the mismatch is largest.
The order goes to the receiving account with its stop loss and take profit converted into that broker’s price precision — a five-digit broker and a three-digit broker do not express the same stop the same way.
Critically, the copier stores the mapping between the master ticket and each slave ticket. Without that record, it cannot later tell which position to modify or close, and a restart turns every copied position into an orphan it will not touch.
Opening is the easy half. The copier must also mirror stop-loss and take-profit modifications, partial closes, full closes and — where the platform allows it — pending order edits and expirations.
It also has to survive interruption. If the connection drops for ninety seconds, a good copier reconciles on reconnect: it compares what the master holds against what each slave holds and resolves the difference. A poor one resumes from the next event and leaves you holding a position on three accounts and no position on the fourth.
The step nobody demos: step five. Any copier looks flawless opening a trade in a sales video. The ones worth paying for are separated by what happens after a dropped connection, a partial fill, or a slave account that hit its margin ceiling while the others did not. Ask about reconciliation before you ask about latency.
Marketing pages describe features. Underneath, there are only three ways to build this, and the choice determines nearly everything else about the product.
A pair of expert advisors: a sender attached to a chart on the master terminal, a receiver on each slave. They communicate through shared files, named pipes, memory-mapped files or a small local socket.
Runs entirely inside MetaTrader. Cheap, widely available, and constrained by everything the MQL sandbox is constrained by.
A separate application outside the terminal that attaches to each platform through its own integration point — the MT4/MT5 manager or gateway interfaces, cTrader’s Open API, or a broker’s FIX session.
Heavier to build, and the only approach that copies cleanly across platforms rather than within one.
A web service. You register accounts on the vendor’s site, their servers hold the logins and connect to the brokers on your behalf, and the copying happens in their data centre.
Nothing to install and nothing to keep running. The trade-off is that a third party now holds working credentials to your accounts.
These are not mutually exclusive labels so much as two independent questions. What does the copier attach to — an EA inside the terminal, or a bridge outside it? And where does it run — your hardware, or the vendor’s? The next two sections take those questions one at a time, because they get conflated constantly and they have entirely different consequences.
This decides the ceiling on speed, reliability and cross-platform reach — and it is invisible from the outside until something fails.
An EA-based copier lives inside MetaTrader. The sender is an expert advisor dropped onto a chart on the master terminal; the receiver is another EA on each slave. They pass messages through the filesystem, a local socket, or a DLL. Everything the copier can do is something MQL can do, and everything MQL cannot do, the copier cannot do either.
An API-bridge copier is a separate program that talks to each platform through a real integration interface — MetaTrader’s gateway and manager APIs, cTrader’s Open API, or FIX where the broker offers it. It does not need a chart, does not run inside the terminal, and is not bound by the terminal’s execution model.
| EA / script copier | Native API bridge | |
|---|---|---|
| Where it runs | Inside the MetaTrader terminal | A separate application beside it |
| Trigger | Ticks, or a timer workaround | Platform events, as they happen |
| Cross-platform (MT4 ↔ MT5 ↔ cTrader) | Effectively no | Yes |
| Charts required | One per instrument, kept open | None |
| Survives a terminal restart | Reattach required | Reconnects on its own |
| Setup effort | Copy two files to a folder | Run a signed installer |
| Typical price model | One-time licence, often cheap | Subscription |
| Best fit | One MT4 pair, one machine, low stakes | Several accounts, mixed platforms, funded capital |
A fair word for EAs: if you are copying one MT4 account to one other MT4 account on the same PC, an EA copier will do it and will cost you almost nothing. The argument for a bridge starts when the account count rises, the platforms stop matching, or the money stops being yours to lose casually.
Where the copier runs is not an implementation detail. It determines who holds your broker credentials, how far your orders travel, who is responsible when it stops, and what happens to you if the vendor has a bad day.
Every copier has to be authenticated against your accounts. To place an order on an account, something must be logged into it. There is no architecture that avoids this. The only real question is where that something lives — and the two available answers produce genuinely different products.
The vendor’s servers hold your logins and do the copying. You configure it in a browser and never think about uptime.
Two network trips, and a third party sitting in the middle holding credentials for both ends.
The copy engine runs on your PC or a VPS you rent. Credentials never leave that machine — and neither does the order path.
The copy happens between processes on one machine. Only each terminal’s own connection to its broker leaves it.
This is the difference that matters most and gets discussed least, so it is worth being precise about what a cloud copier necessarily has.
To place an order on your slave account, the vendor’s server logs into that account. To do that at 3am while you are asleep, it must hold a credential that works at 3am — not a one-time token you approved, but a standing login. On MetaTrader that is an investor or master password; on cTrader it is an OAuth grant or a stored login. Whatever the mechanism, the practical position is the same: a company you have never met can authenticate as you, on all of your accounts, at any time, until you change the password.
Most of these companies are honest and some are competent. That is not the issue. The issue is that you cannot verify either claim, and the aggregate database — thousands of traders, tens of thousands of live logins — is a target whose value has nothing to do with the vendor’s intentions. Security questions to put to any hosted copier, in order:
A self-hosted copier makes all five questions moot, because it never receives the credential in the first place. Passwords sit in the operating system’s own keystore — Windows Credential Manager, macOS Keychain — on the machine running the terminal.
Copier marketing quotes millisecond figures with more confidence than the physics deserves. Here is the shape of it. In a self-hosted setup, master and slave terminals sit on one machine, so the copy itself is a local hand-off — commonly in the tens of milliseconds, and the meaningful remaining delay is each terminal’s own round trip to its broker.
In a cloud setup, the same copy involves your master’s broker reporting the fill, the vendor receiving it, the vendor deciding and dispatching, and then the slave broker executing. That is at least two additional wide-area trips plus whatever queueing the vendor has under load — and the load spikes exactly when everyone’s strategies fire at once, on the news release you cared about.
The honest caveat: for most trading styles this difference is not the thing costing you money. Swing and position trading will not notice. Broker-side execution quality and the spread difference between two brokers usually dominate. It matters for scalping, for news trading, and for anything on an instrument that gaps — which is a real minority of traders, not all of them.
Cloud copying transfers the operational burden to the vendor. That is a genuine benefit and the main reason the model exists. But it also transfers control: when their service degrades at 14:30 on a Friday, your options are to watch the status page and to place your trades manually. Self-hosting keeps both the burden and the control — a copier on a VPS you rent fails for reasons you can inspect and fix.
Neither is strictly safer. A well-run cloud has better uptime than a laptop; a well-chosen VPS has better uptime than most small vendors. The question is not which is more reliable in the abstract, it is which failure you would rather be able to do something about.
| Cloud-hosted copier | Self-hosted copier | |
|---|---|---|
| Where the copy engine runs | The vendor’s data centre | Your PC or your own VPS |
| Who holds your broker password | The vendor, indefinitely | Nobody but your machine |
| Master → slave path | Two internet hops via the vendor | Local hand-off between processes |
| Typical added latency | Vendor round trip plus queueing | Effectively none beyond broker execution |
| Exposure if the vendor is breached | Live trading logins exposed | No credentials to expose |
| Runs while your hardware is off | Yes | No — this is what the VPS is for |
| Who owns uptime | The vendor | You and your host |
| If the vendor shuts down | Copying stops that day | The installed software keeps running |
| Setup effort | Paste credentials into a web form | Run an installer, detect terminals |
| Ongoing cost | Subscription only | Subscription plus ~$10–40/mo VPS |
| Manage from a phone | Yes, natively | Via a dashboard, if the product has one |
| Good fit for funded prop accounts | Only if you accept the credential risk | Yes — the usual reason people move |
The fair summary: cloud copiers trade custody of your credentials for freedom from operations. That is a real trade, not a scam, and for a small account it can be the right one. It stops being the right one at roughly the point where the accounts being copied are large enough, or borrowed enough, that you would not want a stranger able to log into them.
R4Copier is self-hosted, so treat this page as informed rather than neutral — but the ledger above is the one we would give you if you decided to buy someone else’s product. If you want the architecture argument in full, it continues on the self-hosted page and the credential page.
Every copier claims to be fast, reliable and secure. These are the questions whose answers actually vary between products.
Advertised latency numbers. They are measured under conditions nobody discloses and are not comparable between vendors. The architecture tells you more than the number does: a copy that never leaves your machine cannot be slower than one that crosses the internet twice.
Long feature lists. A copier that reliably does five things — detect, map, size, place, reconcile — beats one advertising forty features and dropping a leg after a reconnect. Reliability under failure is not a feature that lists well, which is exactly why it gets skipped.
None of these throw an error. They just quietly produce the wrong positions, which is worse.
The master trades EURUSD, the slave broker expects EURUSD.m, and the copier places nothing — or, if it is being clever about it, places on the wrong instrument. Configure prefixes, suffixes and per-symbol overrides before the first live trade, and verify with a minimum-lot test on each receiving account.
A hedging master can hold long and short on one instrument simultaneously. A netting slave cannot — it merges them into one net position. Copy from hedging to netting and the second order closes part of the first instead of opening beside it. Nothing errors; the account is simply not what you think it is.
This is covered in depth on the MT5 trade copier page.
A fixed multiplier is fine on day one and wrong by month three, because the accounts no longer have the balances you configured it against. On a funded account this is how a drawdown limit gets breached by a position that looked reasonable on the master.
Minimum stop distances, maximum lot sizes, instrument availability and weekend trading hours all differ between brokers. A trade that is valid on the master can be rejected outright on the slave, leaving you hedged on one account and flat on another.
A Windows update reboots the VPS. The terminal reopens without algo trading enabled. The broker forces a re-login after a password rotation. Everything looks normal in the dashboard because the dashboard is reporting the last thing it saw. Alerting on absence of copying is worth more than any latency feature.
Copying to accounts owned by other people can constitute managing money on their behalf, which is regulated in most jurisdictions and typically requires a licence or a specific broker-side structure. Prop firms separately impose their own rules on copying between funded accounts. Read the terms of every account involved; this guide is not legal advice.
Terms used across this guide and on almost every copier's documentation, with the synonyms vendors use interchangeably.
A trade copier is software that watches one trading account and reproduces its orders on one or more other accounts, automatically and within milliseconds.
The account being watched is the master. The accounts receiving the orders are the slaves, or followers. The copier does not decide what to trade — it only mirrors decisions that have already been made on the master, scaled to the size each receiving account should use.
Copying trades between accounts you personally own is a normal, permitted use of trading software with essentially every retail broker. It is the same activity as placing the order twice by hand, executed faster.
Copying to accounts owned by other people is a different activity. Depending on where you and they are located, that can be regulated money management and may require a licence or a specific account structure such as a PAMM or MAM. Prop firms also each have their own rules about copying between funded accounts. Read the terms of every account involved before you connect it, and take professional advice if you are managing other people’s money.
The honest answer is that it depends on the architecture, not on the brand. A local copier, where master and slave terminals run on the same machine, typically passes an order between them in tens of milliseconds because the hand-off is between two processes rather than across the internet.
A cloud copier adds two network trips — terminal to vendor, vendor to terminal — plus queueing at the vendor. That is usually still fast enough for swing and intraday trading and usually not fast enough for scalping on volatile instruments.
Either way, copier latency is only part of the slippage you will actually see. Broker execution speed, spread differences between the two brokers, and the size of the order matter at least as much.
A signal service sends you an instruction — an alert, a message, a subscription feed — and something still has to act on it. A trade copier is the thing that acts, on your own account, without a human in the loop.
The two are often bundled, which is where the confusion comes from. MetaTrader’s built-in Signals marketplace, for instance, is a signal directory with a copier attached to it.
Only if you need it running when you are not. A copier can only copy while it is switched on and connected, so a desktop that sleeps overnight or reboots for a Windows update will miss trades placed in those windows.
If you trade at the desk and close down afterwards, a normal PC is fine. If you are copying to funded accounts around the clock, a VPS is the standard answer and costs less per month than most copier licences.
Across brokers, yes — that is the main reason people buy one. Across platforms depends on the product. Copying MT4 to MT4 is universally supported; copying MT4 to cTrader, or MT5 to MT4, needs a copier that speaks both platforms natively.
Cross-broker copying is also where symbol mapping becomes essential, because two brokers rarely name the same instrument identically. See the section on what silently breaks copying below.
Retail copiers generally run from about $10 to $100 per month, or $100 to $500 as a one-time licence for EA-based products. Pricing usually scales with the number of accounts rather than the number of trades.
Budget for the infrastructure too. A VPS suitable for 24/7 copying is roughly $10–$40 per month, and a copier running on hardware that is not up is worth nothing regardless of what the licence cost.
That depends entirely on where the copier runs, and it is the single most consequential difference between the two architectures.
A cloud copier has to hold a working login for every account, because its servers are the thing logging in on your behalf. A self-hosted copier attaches to terminals already running on your own machine, so the authentication happens locally between two processes and no credential has to be transmitted or stored by a vendor.