I Got Tired of Repricing Binance P2P Ads Manually – So I Automated It
For a while, managing P2P advertisements manually didn’t seem like a problem.
Open the marketplace, check a few competing offers, adjust the price, and continue working. If the market moved later, repeat the process.
With one advertisement, this routine was manageable.
Then I started working with more payment methods and several BUY and SELL advertisements. What had previously taken a few minutes gradually became something I was checking throughout the day.
The problem wasn’t that changing an advertisement price was difficult. The problem was how often the decision had to be made.
The Market Doesn’t Wait
P2P markets can change surprisingly quickly.
A competitor moves an offer. Another merchant disappears because their available volume is exhausted. Someone enters the market with a more aggressive price. A previously irrelevant advertisement suddenly becomes the first visible offer.
You adjust your price.
Twenty minutes later, the situation has changed again.
At some point I realized that I wasn’t really spending my time trading. I was spending it watching other traders.
That distinction matters.
The actual decision I was making was usually quite simple:
- find relevant competing advertisements;
- ignore offers that don’t match my conditions;
- identify the price I want to follow;
- apply a small offset;
- make sure the resulting price remains inside my acceptable range.
The decision itself could be expressed as rules.
If something can consistently be expressed as rules, it is usually a good candidate for automation.
The First Mistake: Following the First Price
My initial assumption was that automation simply meant finding the best visible price and following it.
That turned out to be too simplistic.
Imagine a competitor whose price looks attractive but whose order limits are completely different from yours. Or an advertisement using a payment method you don’t accept. Another merchant might have very little available volume.
Technically, all of these advertisements exist in the same marketplace.
Strategically, they aren’t necessarily your competitors.
So a useful pricing engine needs more than:
Find position #1 → change my price.
It needs filters.
The system should first determine which advertisements are relevant and only then use those offers when calculating a target price.
That changed the way I thought about P2P automation.
Automation Needs Boundaries
There was another problem.
Suppose several automated traders continuously compete with each other. Trader A moves slightly below Trader B. Trader B reacts. Trader A reacts again.
Congratulations: both systems are working perfectly.
Their margins may not be.
An automated pricing system therefore needs limits just as much as it needs speed.
For example, a strategy can define a minimum acceptable price, a maximum price, a required spread or another threshold beyond which repricing simply stops.
That makes the automation much less exciting — and much more useful.
The objective isn’t to win every price movement.
The objective is to execute your strategy consistently.
From Constant Monitoring to Exception Handling
Once I began looking at the problem this way, the role of software became much clearer.
The trader should decide the strategy.
Software should handle repetitive execution.
Instead of manually checking the marketplace dozens of times, the application can monitor relevant advertisements, apply filters, calculate a target price and update selected ads when the configured conditions are satisfied.
One implementation of this idea is the self-hosted P2P Trade Bot from PyDev. What interested me about this approach is not simply automatic repricing, but the idea of keeping the pricing logic and deployment under the operator’s control rather than handing the entire process to a cloud service.
That distinction becomes important when automation is part of an actual trading workflow rather than an occasional convenience.
Self-Hosted vs Cloud Automation
A cloud platform can be convenient because there is nothing to install.
But self-hosted software has a different advantage.
It can run on your own computer or VPS, while configuration and API credentials remain within infrastructure you control.
For technically experienced traders, that also makes customization easier. You aren’t limited to whatever workflow a SaaS dashboard happens to provide.
There is obviously a trade-off: you are responsible for running the application and maintaining the environment.
For me, that’s acceptable.
A small Linux VPS is enough for many automation workloads and can stay online continuously without leaving a desktop computer running all day.
What Automation Actually Changed
The biggest improvement wasn’t that I could suddenly predict the market.
I couldn’t.
Automation didn’t discover a secret trading strategy either.
What changed was much more mundane: repetitive decisions stopped requiring repetitive manual actions.
I could define the conditions once and let software continuously execute them.
If the strategy was wrong, automation wouldn’t magically fix it.
But if the strategy was reasonable, the software could follow it far more consistently than someone checking the marketplace whenever they happened to remember.
And that may be the most useful way to think about trading automation.
Not as a machine that replaces the trader.
As a machine that stops the trader from doing the same boring task 100 times a day.
