There are two ways to build the execution half of a signal copier. You rent an execution API from a company that already runs the plumbing, or you run the trading terminals yourself. The first takes weeks. The second takes months and never entirely stops taking work. We took the second.
Here is the shape of what CopySignals actually runs. One MT4 or MT5 terminal per connected trading account, logged into that account with the credentials the trader supplied, hosted on our infrastructure. Inside each terminal sits a small bridge that accepts one instruction at a time and reports the account state and its open positions back. The parser hands it a trade, the bridge places it, the result comes back with whatever the broker said.
Why not rent the API
A rented execution API is a fine engineering decision for a lot of products, and it is the wrong one for this product. Three reasons, in order of how much they hurt.
The first is diagnosis. When a client tells you their trade did not appear, you need to know whether the parser skipped it, the bridge never sent it, or the broker refused it. With your own terminal you can look at all three. With a vendor in the middle you can look at the request you made and the code you got back, and then you open a support ticket on your client's behalf and wait. A copier is judged on exactly these moments.
The second is coverage. Vendors support the brokers they support. Retail traders following signal channels use small brokers with unusual server names, symbol conventions and account types, and the list of what a vendor accepts becomes the list of clients you can serve. Running terminals means the question is whether MetaTrader can log into that server, which it usually can.
The third is pricing structure. When execution is rented per account, every account you add carries a floor you did not set, and your pricing is written by somebody else's rate card. The broader version of this argument, in a different context, is in platform vendor risk.
What owning it costs
Honesty about the other side. Running terminals is operational work that never finishes. Terminals need to stay logged in, and brokers disconnect sessions. Platform updates arrive on the broker's schedule, not yours. Each terminal consumes resources, so capacity is a real number you have to plan rather than a line item you buy. And a machine that runs somebody's live trading account has to be supervised continuously, because a terminal that is quietly logged out looks exactly like a channel that posted no signals.
Traders who have run their own copier on a VPS know this shape already, described in the forex VPS guide. We took that job on so the trader does not have it, which also means the trader's computer being off changes nothing.
Running the terminals does not make execution instant or guaranteed. Your order still travels to your broker's server and is still subject to their spread, their fills and their rejections. What the design changes is whether anyone can tell you what happened.
One instruction at a time
The bridge deliberately handles a single instruction at a time rather than firing everything at once. It looks like a performance limitation and it is a correctness decision. Order placement, partial closes at each target, moving a stop to break even and a trailing adjustment are all writes against the same account, and a race between two of them produces a position nobody intended. Serialising them means the account state you read is the state you are acting on.
The same restraint applies to what the bridge reports. The account and its open positions come back from the terminal itself rather than being assumed from what was sent, so a rejection or a partial fill shows up as what it was.
Symbols are a terminal problem
Gold is XAUUSD at one broker, GOLD at another, XAUUSD.pro at a third, and XAUUSDm on a cent account. The only authoritative list of what your account can trade is the one your own terminal exposes, so that is what symbol matching reads, with prefix, suffix and rename rules for cases no rule catches. An instrument that does not exist gets a refusal with the reason attached rather than a hopeful near match. This is much harder to do well through an intermediary that has normalised the symbol list before you see it.
What supervision means in practice
The unglamorous half of this decision is monitoring. A terminal that has silently logged out produces no errors, no orders and no complaints, and from the trader's side it looks exactly like a week when the channel posted nothing. So the terminals are watched for whether they are still connected and still reporting the account, rather than only for whether the last order succeeded. Absence of activity is the failure mode that has to be detected, and it is the one a rented interface is least likely to tell you about.
Where the choice shows up for the trader
Mostly in unremarkable ways. The broker list is broader. A refused order comes back with the broker's own reason. A support question has an answer instead of a ticket. And the thing the trader never sees is the point: nobody between us and the broker can change their terms, deprecate an endpoint or go out of business and take your execution with them.
The related decision, and the one people ask about more often, is covered in running without a third party trade API. Trading on margin carries a high risk of loss, and none of this architecture makes a signal a good one.
"When a client says the trade did not appear, you either own the whole path or you own a support ticket. We wanted to own the path."
— Alex Onta, Executive Director, SINGUARD
Key Takeaways
- CopySignals runs one MT4 or MT5 terminal per connected account, with a bridge inside it that takes a single instruction at a time and reports positions back.
- Owning the terminal means a missing trade can be traced to the parser, the bridge or the broker, instead of ending in a vendor support queue.
- Broker coverage becomes a question of whether MetaTrader can log into the server, rather than a question of what an execution vendor supports.
- The cost is continuous operations: sessions drop, platform updates arrive on the broker's schedule, and capacity has to be planned rather than purchased.
Frequently Asked Questions
Do I need to keep my computer on?
No. The Telegram connection and the trading terminals both run on our servers, so nothing depends on your machine being awake.
Does running your own terminals make execution faster?
It removes one hop, but the dominant delays are Telegram delivering the message and your broker accepting the order. The real gain is that every step is visible when something goes wrong.
Which credentials does the terminal need?
The ones a terminal uses to log in to your account, the same ones you would type into MT4 or MT5 yourself. An investor password is read only and cannot place trades, so it will not work for copying.
About the Author
Alex Onta is an Executive Director at SINGUARD. He built eTrader, the terminal, the mobile apps, eTrader Broker, Copytrading, Business and Community, along with the worldwide clustered-server infrastructure it all runs on, with his brother Roman Onta helping on the design, and he leads that division today. Together with Roman he builds the Prop Firm CRM, the Broker CRM, Scalegram and CopySignals, and the two of them carry worldwide compliance, payment processing and international business structuring side by side. He lives and works in Dubai for most of the year. Meet the executive duo leading Singuard's five divisions.