Head to head · checked 2026-09-01
Apirone vs BTCPay Server
Two non-custodial ways to accept Bitcoin without paying to accept it. Apirone charges nothing on acceptance and a fixed fee on withdrawal, with plugins for four platforms including VirtueMart. BTCPay Server charges nothing anywhere, runs on your own server, and names twenty-three integrations.
How do Apirone and BTCPay Server differ?
Weights| Gateway | Score | From | Assets | Verification | Settles fiat | Discloses |
|---|---|---|---|---|---|---|
| | 6.1 | 0% | Not disclosed | 58% | ||
| | 7.7 | 0% | None stated | No | 74% |
Scoring Apirone against BTCPay Server
| Criterion | Weight | Apirone | BTCPay Server |
|---|---|---|---|
| Assets, networks and geography How many assets and chains are live for merchants, and in which markets settlement actually works. | 25 | 4 | 4 |
| All-in cost Processing percentage, conversion spread and payout cost taken together, not the headline number alone. | 25 | 8 | 10 |
| Onboarding and verification What a merchant must submit before going live, how long it takes, and which verticals are refused outright. | 20 | 7 | 10 |
| Integrations and API Maintained plugins, API surface, webhook reliability and whether a developer can reach a test transaction from public docs alone. | 18 | 6 | 9 |
| Support and operations Reachable channels, published response commitments, and documented handling of underpayment and wrong-network sends. | 12 | 5 | 5 |
| Weighted total | 6.1 | 7.7 |
Where the two diverge
Scored 0–10 · checked 2026-09-01
Apirone BTCPay Server
On this page
Free to accept on both, and only one is free to leave#
Apirone charges nothing when a payment arrives and a fixed amount when funds move out. BTCPay Server charges nothing at either point, because there is no company in the path to charge anything.
The fixed withdrawal fee is the entire commercial difference, and its behaviour is worth understanding: it costs the same on a small withdrawal as on a large one, so it punishes frequent sweeps and rewards patience.
Against that, the self-hosted option replaces the fee with a server and somebody's attention. Below a certain volume that is the more expensive arrangement, not the cheaper one.
A service you call against software you run#
One is an Estonian company trading since 2017 that operates the infrastructure for you. The other is an open-source project that states plainly it is not a company.
Both settle to addresses the merchant controls, so the custody property that usually separates providers is identical here. What differs is who is awake when something stops confirming.
That is the real question on this page and it has nothing to do with cryptocurrency. It is a staffing question.
Twenty-three integrations against four#
WooCommerce, Magento, OpenCart and VirtueMart on the hosted side; twenty-three named platforms on the self-hosted side, from WooCommerce and Shopify down to Ghost and Invoice Ninja.
VirtueMart is the interesting entry and it appears nowhere else in this index, which makes the hosted card the only supported option for a Joomla store.
Outside that case the self-hosted project has already done more integration work than any commercial provider here, because each plugin exists for somebody who personally ran that platform.
Bitcoin at the centre of both#
One names Bitcoin, Litecoin and Dogecoin and publishes no count. The other names Bitcoin and Lightning and publishes no count either, with altcoins arriving through separate plugins.
Neither supports a stablecoin, which removes both from most current online checkouts. That is the shared constraint and it is larger than anything else on this page.
Lightning appears on one side only, and it is the difference between a viable and an unviable small payment. For anything under about ten euros, that line decides the pair.
Neither settles fiat#
Fiat settlement is documented as absent on one card and recorded as unpublished on the other. In practice a merchant on either receives crypto and converts elsewhere.
That follows from the shared non-custodial design rather than from any decision either could reverse. A system that never holds a balance does not push bank transfers on a schedule.
A business needing euros in an account should treat this pair as informational and start from the custodial providers instead.
Verification and who stands behind each#
The self-hosted project has no verification because it has no counterparty. The hosted service records merchant verification as undisclosed after checking, which is a question rather than a promise.
On the corporate side the positions reverse: one names an Estonian operator and a founding year, the other names nobody because there is nobody to name.
For a merchant that trade is unusually clean. You are choosing between a company you can identify and no company at all, and both answers are defensible.
Point of sale and invoicing#
Hosted invoicing is documented on both. Point of sale is documented on the self-hosted project and recorded as unpublished on the service.
For a physical business that is a real line, and it runs against the general expectation that the hosted option carries more product surface.
Recurring billing and mass payouts are unpublished on both, so a subscription business or a platform paying out is not served by either card.
How the totals land#
Cost and onboarding both score at the maximum on the self-hosted side, because zero is zero and there is no application to be refused. Integrations score nine against six.
Coverage drags both, at four each, which is what a general scale does to two deliberately narrow products. Neither number should be read as a criticism of fit.
The gap in totals is wide and slightly misleading. It measures what each publishes rather than what running either actually costs a particular business.
What to ask each of them#
To the hosted service: the enumerated asset list and the exact fixed withdrawal amount, since neither is on the public pages and both decide whether the free acceptance is real.
To the project: nothing, but read the deployment documentation before committing. The install is well documented and the ongoing operation is the part people underestimate.
To yourself, on both: who holds the wallet keys in eighteen months, and whether the recovery phrase has ever been restored on a spare device.
Where this pair actually lands#
A Joomla or Magento store taking Bitcoin, withdrawing occasionally, takes the hosted service and pays a fixed amount a few times a year with nobody to employ.
A business with an engineer and a Lightning use case takes the project, pays nothing at all, and accepts a commitment measured in years.
A shop that wants free acceptance and has nobody to run a server should take the hosted card and stop reading. The cheapest option here is only cheap for somebody who can spare the time.
Read next
Questions merchants ask
Are both actually free to accept payments?
Yes on published terms. Apirone charges a fixed fee on withdrawal instead of on acceptance; BTCPay Server charges nothing at any point, and its cost is the server and the person who maintains it.
Which one supports Lightning?
BTCPay Server names Bitcoin and Lightning. Apirone names Bitcoin, Litecoin and Dogecoin without Lightning, so its smallest viable payment is set by on-chain network fees.
Is Apirone or BTCPay Server cheaper?
Neither publishes enough for a like-for-like answer. BTCPay Server states no processing rate at all, so this comparison has to happen over email.
Can you run Apirone and BTCPay Server at once?
Yes, and during a migration you should: a billing cycle with both live tells you more about settlement timing and failure handling than any published page. Both name WooCommerce and Magento among their integrations, so the two can sit side by side at the same checkout without a second build.