Understanding what you pay

The price of a conversion is made of at most four elements, and only one of them is our revenue. This guide describes each line, how it is computed, and how to check that the total adds up — because a breakdown you cannot verify is not a breakdown, it is a claim.

The principle: the lines sum to the gap

The pricing engine's design rule is short:

The sum of the fee lines must equal exactly the difference between the value at mid-market and the net received. If the equality does not hold, the quote is refused.

That check runs on every quote, before display. So there is no "miscellaneous fee", no favourable rounding, no residual gap absorbed somewhere. Either the lines add up, or nothing is shown.

It is a technical constraint, not a marketing promise. It has a cost: it forbids folding a margin into a rate, which is the industry's most widespread practice.

Line 1: the mid-market rate

Mid-market is the midpoint between the best bid and the best ask on reference venues. It is not an executable rate: nobody trades in the middle of the book. It is a shared reference point.

Two pieces of information are needed for a mid-market rate to mean anything:

• Its source. Which market data, aggregated how.
• Its timestamp. A three-minute-old rate is not a rate.

Both are published with every quote and on the live rates page. A rate with no source and no timestamp binds nobody, and it is the first place a cost can hide: announcing "the market rate" from a reference already shifted by 0.3% is worth 300 EUR on 100,000, invisible to the naked eye.

Line 2: the spread

This is our revenue. It is expressed in basis points — one basis point is one hundredth of a percent — and computed on the gross value at mid-market.

The spread applied combines several components:

• A base spread, identical for everyone.
• A volatility surcharge for non-stable assets. While your deposit waits for confirmations the price moves. On a stablecoin that risk is structurally smaller, and the surcharge does not apply.
• A beta corridor surcharge for asset/rail pairs whose operational failure rate is still poorly known.
• Volume tier discounts, applied automatically above certain amounts. There is no individual negotiation: the schedule is public.

The values in force are published on the fee schedule. They are what the engine applies, not an indicative range.

Line 3: the network cost

This is an estimate of the on-chain cost associated with your deposit. It depends on the network, not on the amount: moving 100 USDT or 10,000 USDT on Tron costs the same.

Direct consequence: this line is negligible on a large amount and significant on a small one. That is the main reason for per-asset minimum deposits — below them, the network cost line eats an absurd share of the operation. The full comparison is in USDT on TRC-20 or ERC-20.

This line does not cover the fee you pay your origin platform to withdraw. That one does not pass through us and does not appear in our breakdown — but it counts in your total cost, and you must add it to compare honestly.

Line 4: the rail fee

Each payout rail has its own tariff: a fixed part, a part in basis points, or both. A SEPA Instant transfer costs a few cents; a SWIFT wire costs tens of euros and involves intermediary banks taking their share along the way.

That fee is published on every rail page before you choose, not at the moment you confirm.

The line hidden elsewhere: FX

When the payout currency is not the one the asset is valued in, a conversion happens. Receiving reais, naira or rupees means somebody sourced those currencies, at a rate.

It is often the heaviest line on emerging corridors, and the one most frequently folded into the headline rate. Our rule is the same as for the spread: it appears separately, with its calculation basis.

Checking it yourself, in one division

Verification requires no tools.

effective rate = net received / quantity sent
total cost     = (mid-market - effective rate) / mid-market

The result, as a percentage or in basis points, absorbs everything: spread, network cost, rail fee, FX. It is the only comparable value between two providers, whatever way each presents its lines.

We show that number directly on every quote, as the gap to mid-market. A service that does not show it leaves you to compute it; a service that refuses to give an exact net before you send has not given you a price.

A complete example

Selling 0.25 BTC, mid-market at 96,000 EUR per BTC, payout over SEPA Instant. Illustrative values.

| Line | Basis | Amount | | --- | --- | --- | | Value at mid-market | 0.25 x 96,000 | 24,000.00 EUR | | Spread | basis points on gross | -246.00 EUR | | Network cost | estimated Bitcoin withdrawal | -1.80 EUR | | Rail fee | fixed SEPA Instant | -0.35 EUR | | Net received | | 23,751.85 EUR |

Check: 246.00 + 1.80 + 0.35 = 248.15, and 24,000.00 - 248.15 = 23,751.85. The equality holds.

Effective rate: 95,007.40 EUR per BTC. Gap to mid-market: 103 basis points.

Why we do not advertise "zero fees"

The "zero fees" claim is technically true for almost everyone who uses it, and that is what makes it effective.

The mechanism is simple. The service charges no line labelled "fee"; it applies a rate below mid-market. The margin is identical, it has just changed name. On a 10,000 EUR sale, a 1.5% gap on the rate is 150 EUR — exactly what a 150 EUR fee line would have cost, without appearing anywhere.

Three reasons we do not make that choice.

It makes comparison impossible. Two "zero fee" offers can differ by two percentage points. Without a breakdown, nothing lets you see that before the money arrives.

It prevents verification. A customer cannot redo a calculation whose inputs they do not know. A breakdown that cannot be checked is a claim.

It moves the conversation. When the margin sits in the rate, the question becomes "what rate do you show" — which can always be answered with a flattering, undated number. When it sits on a line, the question becomes "how much do you take", which has one answer.

The cost of that stance is real: showing a spread line looks more expensive than showing zero, even when the net received is higher. We accept it, because the alternative is being cheaper in appearance and more expensive in practice.

The questions to ask any service

1. What is the exact net for this precise amount, right now?
2. For how long is that net firm?
3. What is the source of the reference rate, and how old is it?
4. Which lines make up the gap between gross and net?
5. Is there a currency conversion, and at what rate?

A service that answers all five with numbers has given you a price. A service that answers "our fees are among the lowest on the market" has given you an argument.

The cost that appears on no invoice

Two real costs never show up as a line, and both are worth counting.

The first is time. A rate two basis points better, settling three business days later, leaves you exposed to the market for three days if you need to convert back. On a volatile asset that exposure can dwarf the fee difference many times over.

The second is failure. A payout rejected for a name mismatch or a closed account is not lost, but it takes days to bounce and requires a fresh transfer. A provider that fails one payout in twenty is more expensive than a slightly pricier one that validates your identifier, resolves the account holder name and refuses a malformed entry before you confirm.

Neither cost is hypothetical, and neither belongs in a fee schedule. They belong in the comparison you make before choosing.

Articles attached to this guide

• Headline rate versus effective rate develops the comparison method and the most common traps.
• USDT on TRC-20 or ERC-20 shows the relative weight of network cost by amount.
• Underpaying an order explains how fixed and variable lines behave when the amount received changes.