The word that costs 282 satoshis

Two hundred Tether on Tron into Bitcoin, asked twice in the same pass. The only difference between the two columns is the network asked for — bitcoin in one, lightning in the other:

Identical to the satoshi, all four. The same held for ln and btcln: three different spellings, one unchanging answer. The network parameter is being dropped on the floor and an on-chain price is coming back wearing a Lightning label.

Nothing in the response says so. There is no warning, no note, no error. A comparison built on those four numbers would show Lightning as available, priced, and no cheaper — three claims, all wrong.

ChangeNOW, SideShift, SimpleSwap and StealthEX returned errors on every spelling. ChangeNOW said the pair was inactive; the others failed on the ticker. None of them quoted anything.

That is the preferable failure. A refusal is visible and can be handled. A wrong number delivered confidently cannot, because nothing downstream knows to doubt it. The same distinction runs through the network names page, where asking for Monero by the word monero rather than xmr halves the field — quietly, with fewer providers answering rather than an error explaining why.

Lightning is not a network you request alongside Bitcoin. To the one provider that handles it, it is its own ticker. Asking for btcln as the currency rather than btc on a Lightning network changes the answer — and only FixedFloat answers at all:

The first two rows are the same number because the first two questions are the same question. The third is different because it is finally the right one.

Asked at four sizes across a fiftyfold range, in one pass, the gap between on-chain Bitcoin and the same transfer over Lightning never moved:

Two hundred and eighty-two satoshis, four times, to the satoshi. That is not a better rate. It is a fixed cost that stops being charged — the on-chain withdrawal fee, which Lightning does not incur and which the provider therefore does not pass on.

Because it is fixed, what it is worth depends entirely on the size of what you are moving: fifty times more valuable on twenty dollars than on a thousand. That is the same arithmetic as why small swaps cost more, seen from the other side. The fixed components of a swap are what punish small transfers, and Lightning removes one of them.

Crypto Train cannot route a Lightning payout today, and the reasons are worth stating rather than leaving as an absence.

The coin catalogue carries 320 entries and none of them is Lightning, so it cannot be selected. Address validation resolves a chain-qualified ticker through ChangeNOW — one of the four providers that refuses Lightning outright — so a Lightning invoice has nothing to validate against. And sending from Lightning is not a matter of adding a ticker at all: this app signs a transaction in your browser, and a Lightning payment has no transaction to sign. That would be a different product.

A Lightning payout would therefore mean one provider, a catalogue entry, a validation path that bypasses the usual check, and a verification that the leg genuinely settles off chain rather than silently going on it. That is a real piece of work, and this is a note about what was measured rather than a promise about what comes next.

Every figure here was taken on 30 September 2026 between 17:28 and 17:30 UTC, Tether on Tron into Bitcoin, all eight providers asked in the same pass for each row. Prices move; the tables above are a reading rather than a constant, and the difference that matters — 282 satoshis — is the one that held still across all four sizes.

One limit, stated rather than buried. These requests went through this site’s own comparison layer, not to each provider’s raw API. A provider might accept some other spelling that the layer never sends, in which case it would appear here as a refusal when the truth is that it was asked the wrong way. The finding that four providers return an identical on-chain price for a Lightning request does not depend on that caveat; the four refusals do.