To accept crypto payments, a business does five things: decides which coins it takes and which it keeps, picks a gateway and passes KYB verification, connects a checkout flow, configures settlement and conversion, and sets up accounting, refunds and support. On Speend, KYB takes one to three business days after documents arrive, and a plugin install runs under two hours.
This guide is for merchants already selling online who are deciding whether to add a crypto rail or move provider.
What does it mean to accept crypto payments?
A customer sends a blockchain transaction to an address generated for their order, and your system marks the order paid once the network confirms it. The money arrives on-chain rather than through a card network, so there is no issuer, no acquirer and no dispute mechanism in the middle.
That difference drives almost everything below. There is no chargeback, because no institution can reverse a confirmed transaction. There is no card-on-file, because the buyer signs one transfer rather than granting a standing right to charge them. And what sits on your balance afterwards is a separate setting from what you accept at checkout.
A crypto payment gateway sits between the blockchain and your order system: it generates the address, watches the chain, screens the incoming wallet against AML lists, credits your balance and fires a webhook your backend can act on.
Why do businesses accept crypto in 2026?
Three reasons hold up under scrutiny: cost per transaction, access to buyers whose cards fail, and the absence of chargeback losses. A fourth, “crypto demand is growing”, holds up only in relative terms.
Cost. Stripe’s published US pricing, checked 30 July 2026, is 2.9% plus 30¢ per domestic card transaction, with 1.5% added for international cards and 1% more when currency conversion is required. Speend charges 0.5% mono-currency, passes network fees through at blockchain cost, and takes no setup or subscription fee. The costs section below works the gap through on a real turnover.
Access. Cross-border card acceptance fails for reasons unrelated to the buyer’s solvency: issuer geo-blocks, bank policy on the merchant category, correspondent friction. The same friction shows up in the cost of moving money at all. On the US to Mexico corridor a bank wire runs an estimated $115 to $200 all-in once the sending fee, correspondent deductions and the FX spread are counted, against $10 to $30 on a stablecoin rail. The World Bank’s Remittance Prices Worldwide for Q3 2025 put the consumer-side International MTO Index at 5.52%.
Chargebacks. An on-chain payment is final once confirmed, so friendly fraud — the buyer who receives goods and files a dispute anyway — has no mechanism to operate through.
Demand, stated honestly. Industry estimates put merchant acceptance above 25 million businesses worldwide in 2025, with roughly 18,500 on integrated crypto gateways against about 12,000 in 2023. The survey data underneath is firmer. The National Cryptocurrency Association, surveying 619 payment decision-makers with PayPal and The Harris Poll in October 2025, found 39% of US merchants accepting crypto, rising to 50% above $500 million in revenue. On the buyer side, eMarketer forecast US crypto payment users growing 82.1% between 2024 and 2026 — while putting that at 2.6% of the US population and calling it a nascent-stage payment option.
Where it does not pay off: if your buyers are domestic consumers in one card-friendly market, crypto adds operational surface without adding revenue. The rail earns its place when a measurable share of demand is cross-border, high-ticket or already crypto-native.
How do you accept crypto payments in five steps?
Steps one and two can run in parallel: while documents are in review, your developer reads the docs. Step three is the only one that needs developer time, and often not much of it.
Step 1. Decide what you accept and what you keep
Separate two decisions that merchants routinely merge. What the customer may pay with is a conversion question: the wider the choice at checkout, the fewer abandoned payments. What sits on your balance afterwards is a treasury question. A store can accept 300 coins and hold nothing but USDT, because auto-conversion sits between the two. Start narrow on holding — a stablecoin-only treasury removes exchange-rate exposure and makes reconciliation arithmetic instead of judgement.
Step 2. Choose a gateway and pass verification
Gateways run KYB on the merchant: company documents, ownership structure, and a description of what you sell. On Speend that takes one to three business days from the moment documents arrive, with a personal manager working through it alongside you. Compare providers on the four things that actually vary — the published fee and what it includes, settlement options, networks supported, and whether a sandbox exists. You can compare 12 crypto payment gateways on exactly those.
Step 3. Connect the payment flow
Integration takes one of five shapes, covered in the next section; the fastest is a plugin — install it, paste your API key, enable the coins you want. Test in sandbox first. Speend’s sandbox mirrors production one-to-one, so the webhook payloads you handle in testing are the payloads you handle in production.
Step 4. Set up settlement and conversion
Settlement answers one question: in what asset do you want to end up? Hold what arrives and you carry price exposure on volatile coins. Auto-convert to a stablecoin and you hold a dollar-denominated asset with no FX drift. Or convert to fiat, which Speend offers through partner banking rails in selected jurisdictions. Set a withdrawal cadence at the same time: daily caps the balance at risk, weekly reduces network fees on outgoing transfers.
Step 5. Set up accounting, refunds and support
Decide before launch how a refund gets issued, who approves it and from which wallet, and what support tells a customer who paid the wrong amount. The failure modes are covered further down; what matters is that the answers exist before the first incident. Bookkeeping treats an incoming payment as receipt of an asset valued at the transaction moment, not a bank deposit, so your accountant will want the rate source and timestamp on every order — check that your gateway exports both.
Which integration method fits your business?
The choice is driven by where the buyer is and how much developer time you have. Hosted checkout and payment links need no development; a widget needs front-end work; an API integration needs a backend and webhook handling.
| Method | Developer time | Best for | Main constraint |
|---|---|---|---|
| Hosted checkout page | None to minimal | Stores wanting fastest launch | Buyer leaves your domain |
| On-site widget | Low, front-end only | Brand-sensitive checkouts | Front-end work on your side |
| API integration | Days | Custom storefronts, marketplaces | Needs backend and webhook handling |
| Payment links and invoices | None | B2B, freelancers, one-off deals | Manual per-invoice unless automated |
| QR codes in person | None | Retail, events, hospitality | Requires staff process for failures |
Hosted checkout page
The gateway renders and secures the page: you redirect the buyer with an order amount, they choose a coin and network, and return once payment is submitted. No payment data touches your infrastructure, which is why most stores that launch in a day launch this way.
On-site widget
The payment step sits inside your own page, so the buyer never leaves your domain. The price is front-end work, so pick it when abandonment on redirect is a measured problem rather than a suspected one.
API integration
Your backend creates the order, receives the address and listens for webhooks. You control the interface completely and own the failure handling completely, which is the real cost. Marketplaces and subscription products end up here: Speend ties recurring charges to a wallet authorisation the customer can revoke, and SaaS billing has its own rules.
Payment links and invoices
A URL with an amount attached: send it by email, chat or invoice, and the buyer pays with no integration on your side. For B2B and one-off deals this is often the entire implementation, though it scales poorly past a few dozen invoices a month unless you generate links through the API.
QR codes for in-person payments
The code encodes address, amount and network: the buyer scans, their wallet prefills, they confirm. Confirmation speed matters more here than anywhere else because a queue is waiting, so use a fast network at the counter. Retail follows the same pattern as e-commerce.
How long does onboarding take, and what documents do you need?
One to three business days for KYB after documents arrive, plus under two hours for a plugin install or a few days for a custom API build. A store submitting complete documents on Monday can be live by Wednesday. Prepare these first, because incomplete packets cause nearly all delay:
- certificate of incorporation and current company registry extract
- ownership and beneficial-owner structure, with ID documents for the disclosed owners
- a working website or product description showing what you actually sell
- the corporate bank details or settlement wallet you intend to use
- for regulated verticals, the licence covering the activity
Verification applies to your company, not your customers. Screening on the payment side looks at the incoming address and its history; the payer’s identity sits outside that procedure.
When is a payment final, and what happens to the order until then?
An order moves through three states and support questions cluster in the middle one. The gateway generates an address and the order sits unpaid. The customer broadcasts a transaction and the order goes pending: money visible on-chain, not yet credited. The network reaches the required confirmations, the gateway credits your balance and fires the webhook, and the order is paid.
Want to accept crypto payments on your website?
Fast setup and KYC/KYB, fee starts from 0.5%
Contact UsHow long the middle state lasts is a property of the network, not the gateway. TRON produces a block roughly every three seconds and Speend treats USDT on TRC-20 as confirmed in under 30 seconds; Solana confirms in under a second; Ethereum takes one to three minutes and stretches towards half an hour under congestion; Bitcoin is conventionally settled after one to six blocks, at roughly ten minutes each.
Two decisions follow, and both belong in your order logic rather than your support inbox: whether a pending order reserves stock, and what the customer sees while it waits. Fast rails shrink this window; they do not remove it.
What happens when a payment goes wrong?
Four failure modes cover almost everything: wrong amount, wrong network, payment after the quoted rate expired, and refunds. None has a card-network equivalent, which is why they surprise merchants in month one rather than month six.
Underpayment and overpayment
Crypto amounts are entered manually more often than card amounts, and some wallets deduct the network fee from the transfer itself, which produces short payments. Three policies work: accept a tolerance band and mark the order paid, hold the order and request the difference, or refund in full and ask the customer to retry. Pick one before launch, write it into your terms, and give support a script.
Wrong network
The same token exists on several chains, and USDT is the common case. A customer used to sending crypto to another person copies the address and never checks the network, so USDT-ERC20 goes to an address that may not exist on that chain. Recovery ranges from straightforward to impossible.
Prevention is cheaper. Show the network at the same visual weight as the coin, prefill it in the QR payload, and repeat it in the confirmation email. Merchants who take USDT should understand choosing a USDT network before enabling all five.
Late confirmation
Gateways quote a rate for a limited window, so a customer who opens checkout, walks away and pays 40 minutes later returns to an expired quote. Two things reduce this: a visible countdown, and a defined behaviour on expiry that support knows by heart.
Refunds without chargebacks
A refund is a new outbound transaction you initiate, not a reversal, and nobody can force it — an advantage on fraud, a burden on customer service. Three rules make it manageable. Refund only to an address the customer confirms in writing, never to the sending address by assumption, because exchange deposit addresses do not accept returns reliably. Decide who pays the network fee. Refund in the same asset that arrived, so neither side absorbs a rate move.
Which coins should you accept first?
Start with stablecoins on a fast network, add Bitcoin, then widen. USDT and USDC cover most merchant volume and remove volatility from your balance; Bitcoin brings a buyer segment that will not pay in anything else. Everything beyond those three is optimisation. Speend supports 300+ assets across 18 networks, including Bitcoin, USDT, Ethereum, USDC and Litecoin, so widening later is a dashboard toggle rather than a project. Dedicated pages cover USDT, Bitcoin and Ethereum in detail.
The network matters as much as the coin. TRC-20 keeps per-transaction cost under a dollar, which is why it carries most merchant flow, while ERC-20 is worth enabling for B2B customers whose treasury already holds it. Stablecoins and Bitcoin also differ as treasury assets.
What does accepting crypto actually cost?
Three components: the gateway fee, the network fee, and the cost of converting into whatever you actually spend. Speend charges 0.5% mono-currency, passes network fees through at blockchain cost, and takes no setup fee and no subscription. Auto-conversion starts at 0%: converting incoming Bitcoin into USDT, USDC, EUR or USD costs nothing, and conversion on USDT and Ethereum is 1%. The arithmetic on a concrete store, at $200,000 monthly turnover across 2,500 orders:
| Rail | Fee structure | Monthly cost on $200,000 |
|---|---|---|
| Speend, mono-currency | 0.5%, no per-order fee | $1,000 |
| Speend, Bitcoin with auto-conversion | 0.5% + 0% conversion | $1,000 |
| Speend, USDT or ETH with auto-conversion | 1%, no per-order fee | $2,000 |
| Stripe stablecoin payments | 1.5% | $3,000 |
| Cards, domestic (Stripe published) | 2.9% + 30¢ | $6,550 |
| Cards, international (Stripe published) | 4.4% + 30¢ | $9,550 |
Card rates and the 1.5% stablecoin rate are Stripe’s published US pricing as checked on 30 July 2026; Stripe’s stablecoin acceptance is available to US-registered businesses. Percentage fees on both rails apply to the same turnover, so the comparison is like for like, with one caveat that matters.
The caveat. Those numbers assume all $200,000 arrives on the rail in question, and it will not. At 5% crypto share the comparison is $10,000 processed: roughly $327 on domestic cards against $50 on Speend mono-currency. At that share the rail pays for itself through orders that would otherwise fail at an issuer and through dispute costs it never generates — card economics carry chargeback fees and lost dispute value, and a confirmed transaction produces neither. Fee savings become the main argument once crypto share is material or your mix is international.
What is your business responsible for on compliance?
The split is simple once written down, and most confusion comes from never writing it down.
| Task | Gateway | Your business |
|---|---|---|
| Screening incoming funds against AML lists | ✓ | |
| Verifying your company (KYB) | ✓ | |
| Custody of funds before withdrawal | ✓ | |
| Deciding policy on underpayment and refunds | ✓ | |
| Executing refunds | ✓ | |
| Rate source and timestamp for each order | provides the export | reconciles and files |
| Tax treatment and revenue recognition | ✓ | |
| Licensing for your own activity | ✓ | |
| What your checkout discloses to customers | ✓ |
Speend runs on-chain AML screening on every incoming transaction and rejects flagged funds automatically, with MPC wallet custody, role-based access with an audit log, and a 99.95% uptime SLA. Screening runs at receipt rather than at withdrawal, which is where the same risk would otherwise surface months later when you move funds to an exchange. The right-hand column does not transfer to any provider at any price.
FAQ
How long does it take to start accepting crypto payments?
A store submitting a complete document packet at the start of the week is normally live within that week. The long pole is verification, not integration.
Do I need a crypto wallet to accept crypto payments?
No. The gateway generates an address per order and credits a merchant balance you withdraw from. You need a wallet only as the destination for withdrawals, and even that can be a stablecoin balance you leave in place.
Can a small business accept crypto payments?
Yes. Speend charges no setup fee and no subscription, so cost scales with volume rather than company size, and payment links require no integration at all.
How do refunds work with crypto payments?
You send a new outbound transaction to an address the customer confirms in writing. Budget for the operational side: refunds are manual, and someone has to absorb the network fee.
What happens if a customer sends the wrong amount?
The gateway credits what arrived and flags the order rather than resolving it for you. The deciding factor between the three policies is usually order value: a tolerance band that makes sense on a $300 order is expensive on a $30 one.
What happens if a customer pays on the wrong network?
Recovery depends on which chains are involved and is sometimes impossible, so treat this as a checkout design problem rather than a support one. Labelling the network as prominently as the coin removes most cases before they happen.
Do I have to hold cryptocurrency if I accept it?
No. Auto-conversion moves incoming assets into a stablecoin at receipt, from 0% on Bitcoin, and Speend also offers fiat settlement through partner banking rails in selected jurisdictions.
How many confirmations before a payment is final?
USDT on TRC-20 confirms in under 30 seconds and Solana in under a second, while Bitcoin is conventionally treated as settled after one to six blocks at roughly ten minutes each. Your gateway sets the threshold, so the practical question is what the order shows the customer meanwhile.
Get started with Speend
Speend is a crypto payment gateway built for merchants who want the rail working this week rather than next quarter: 0.5% mono-currency, auto-conversion from 0%, 300+ coins across 18 networks, KYB in one to three business days, plugins for WooCommerce, Shopify and Magento plus a REST API, and a personal manager you can reach directly. Verticals differ enough to be worth reading separately: e-commerce and SaaS and subscriptions each have their own integration patterns.
