Ahmad AffanTrading automation
Trading automation: signals to executed trades, with the controls to trust it.
I build the software around trading products: Telegram signal ingestion, parsing and validation, queued execution to MetaTrader, Expert Advisors, licensing and subscription billing.
Automating trading is mostly engineering discipline. Signals arrive in messy human text, brokers differ, networks fail, and every missed or duplicated order costs money. The system has to parse reliably, check risk rules, queue work so spikes never drop a trade, and report exactly what happened.
I built a subscription product that does this end to end: Telegram ingestion, Redis queues, execution through a customer’s own MT4/MT5 EA or a cloud API, Stripe billing with seat limits enforced on the server, and an admin panel to run it.
Telegram signal copiers
Watch chosen channels, parse signals and execute them on the customer’s MT4 or MT5 account.
Custom Expert Advisors
MQL4 and MQL5 EAs with licence checks, trade commands, heartbeats and risk limits.
Licensing and subscriptions
Plans, trading-account limits and licence enforcement for signal and EA products.
Cloud execution
Run trades through cloud APIs for customers who don’t want a VPS.
Trade journals and dashboards
Positions, history and performance views for you or your customers.
Risk and monitoring tools
Kill switches, max-lot and daily-loss rules, heartbeat monitoring and alerts.
Signal ingestion
An MTProto worker watches chats using an encrypted session, so no manual copying is needed.
Parsing and validation
Turn free text into structured signals, then run duplicate and risk checks.
Queueing
Redis and BullMQ queues so bursts never lose a trade and everything is retried safely.
Execution routing
Send trades to the customer’s EA by HTTP polling or to a cloud execution API.
Expert Advisors
MQL4 and MQL5 code with broker-side stop loss and take profit and status heartbeats.
Billing and licensing
Stripe subscriptions, prorated plan changes and server-enforced account limits.
Admin panel
Manage customers, plans and settings without a redeploy.
| API | NestJSTypeScriptNode.js |
|---|---|
| Front end | Next.jsReactTailwind CSS |
| Trading | MQL4 / MQL5MetaTrader 4 and 5Cloud execution APIs |
| Ingest | Telegram MTProto (GramJS) |
| Queues and data | RedisBullMQMongoDBSocket.IO |
| Billing and security | StripeAES-256-GCM encryption |
Define the rules
Which signals, which brokers, what risk limits, and what the system must never do.
Build ingest and parser
Capture and structure signals, tested on historical messages.
Add queue and execution
Route to EA or cloud execution on demo accounts first.
Layer on billing and admin
Plans, limits, licences and dashboards.
Go live with monitoring
Logs of every send, acknowledgement and fill, plus alerts.
Copy delay is not slippage
Delay is time between master and follower orders. Slippage is price difference. They need different fixes, so log both.
Brokers differ
Symbols, lot sizes, execution modes and spreads vary, which is why identical results are never guaranteed.
Python or MQL5?
MQL5 runs inside the terminal and suits execution. Python suits research, data and monitoring, with extra process overhead.
Not financial advice
I build tools, not strategies. Trading is risky and past results do not guarantee future results. Test on demo first.
Work in this area, shipped and live.
Yes. I have built one that ingests Telegram signals, validates them, queues them and executes through an EA or a cloud API.
No. I build the tooling and execution around a strategy you already have. I don’t offer financial advice.
MQL5 for execution inside the terminal, Python for research and monitoring. Many systems use both.
Yes. I add licence checks, plan limits and a customer dashboard, with Stripe handling payments.
With explicit rules: max lot, spread and slippage filters, symbol mapping, daily limits and a kill switch, all logged.
Other services