Trading Bot Fees Explained: What You're Actually Paying Per Trade

Every trading bot puts a percentage on its homepage. One percent, half a percent, whatever it is. And that number is real, it's just nowhere near the full picture of what a trade costs you.

I think most people know this vaguely but have never actually added it up. So here's the whole thing, in order of how much it usually costs you, which is not the order you'd expect.

1. The platform fee

The advertised one. Most of the category sits at 1% flat, some are lower on specific chains or specific order types, and a few run tiered systems where your rate depends on volume.

Two things worth knowing here.

It's charged on both sides. A 1% fee is 1% going in and 1% coming out, so a round trip is 2% before anything else happens. People forget this constantly and then wonder why a 3% gain came out flat.

And watch for conditions. "0.5%" that only applies to manual trades on one specific chain is not a 0.5% bot, it's a 1% bot with a discount you might occasionally qualify for. Same with volume tiers, where the headline rate requires monthly volume most people will never do. I'd rather one number that's true everywhere than a good number with an asterisk, but that's partly why we price the way we do, so take that as you like.

2. Network fees, and the one nobody mentions

Gas is obvious and unavoidable. What's less obvious is the priority fee, which is the extra you pay to get your transaction confirmed faster, and this is where bots quietly differ a lot.

Some bots default to a much higher priority fee than they need to. It makes execution feel snappy, which is good marketing, but on a small trade that fixed cost can be a meaningful percentage of your position. If you're doing 50 dollar trades and the default priority fee is aggressive, you might be paying more in priority than in platform fees without ever seeing it broken out.

This is worth checking on any bot you use. Look at what the priority fee default actually is, and whether you can adjust it. A bot that dynamically adjusts based on network conditions rather than always paying top dollar will save you real money over a lot of trades.

3. Token taxes

Some contracts take a cut on every buy and sell, written into the token itself. Nothing to do with your bot. A 5% sell tax means you're down 5% the moment you're in.

The number matters less than whether it can change. Plenty of tokens launch with a reasonable tax and raise it later, which is one of the ways honeypots work, and I've gone through the rest of them in the rug pull checklist.

You can automate this one. Setting maximum buy and sell tax limits means orders won't fire on tokens that breach them, so you're not relying on remembering to check every time.

4. Slippage and price impact

This is where most people actually lose the most money, and almost nobody counts it as a cost.

If you buy at a 10% worse price than you saw, that's a 10% cost. It's just not itemised anywhere, so it doesn't feel like one. Your position simply starts smaller than you expected and you shrug.

Part of that is your own trade moving the price, especially on thin liquidity. Part of it is the market moving in the gap between clicking and confirming. All of it is real money and it's frequently larger than every other cost on this list combined. There's a full breakdown in what slippage actually does, but the short version is that a lot of people set it once, forget about it, and pay for that for months.

5. MEV extraction

The invisible one. Bots watching the public mempool can front-run your trade, push the price up, let you fill worse, and sell straight after. You'll never see a line item for it. Your fill was just worse than it should have been.

On thin memecoin pools with wide slippage tolerance, this is not rare. It's closer to routine. I've written about how sandwich attacks work separately, but for the purposes of this article, treat it as a variable tax on every unprotected trade you make.

Adding it up

Say you buy 500 dollars of something with a 5% token tax, 8% effective slippage, and a 1% platform fee, then sell at the same price with the same conditions.

You're down roughly 28% on a trade where the price never moved.

That's a deliberately unflattering example, and a well-configured trade on a clean token looks far better than that. But it illustrates the point: the platform fee, the thing everybody compares, was 2 percentage points of a 28 point cost. Everyone argues about the smallest number on the list.

What to actually optimise

In rough order of what saves you the most:

Slippage settings, per trade. Biggest lever by a distance and it's free to fix. Tighter on liquid tokens, only as wide as necessary on new launches.

MEV protection on. Stops your slippage tolerance being an advertisement to bots. Also free.

Tax limits configured once. Filters out an entire category of expensive mistake without you doing anything.

Priority fee defaults. Check what your bot is doing here, particularly if you trade small sizes.

Then the platform fee. It matters, and over hundreds of trades the difference between 0.9% and 1% adds up. But it's the last thing on this list for a reason, and if you're picking a bot purely on the headline rate you're optimising the wrong variable. There's a proper side-by-side in the Telegram trading bot comparison if you want the numbers.

Where we sit

Axxel is 0.9% flat. Every chain, every order type, no subscription, no premium tier, and no volume gate where you need to push millions through before you get the good rate. Whatever you're trading, that's the number.

You can set slippage and maximum buy and sell tax limits in settings, and MEV protection is available on trades across all five supported chains. There's also cashback on your own trading fees, and if you refer people, revenue share on theirs.

None of which changes the main point of this article, which is that your fee schedule is probably not where your money is going. Fix the settings first.

Crypto trading carries risk. Nothing here is financial advice. Axxel is not available in all regions.