Dwolla Review

An API for moving money between US bank accounts, founded in Des Moines in 2008 and one of the earliest companies to make ACH programmable. Dwolla does not touch cards at all: it does standard ACH, Same Day ACH, and instant payments over the RTP network and FedNow, behind a single integration. That is a genuinely useful thing to be able to buy, and Dwolla has been doing it longer than almost anyone. Two things should temper the enthusiasm. It no longer publishes any pricing — the pricing page now says only that terms are tailored to volume, rails and integration — and in May 2026 Dwolla was acquired by NMI, which is folding its capabilities into NMI's embedded payments platform. NMI says it will keep supporting existing Dwolla customers; neither company has said whether the Dwolla brand and product survive the integration. Build on it if bank-to-bank money movement is your requirement, but ask about the roadmap in writing.

Estimated annual cost at three realistic processing volumes, using Dwolla’s published online rate plus monthly fees. Real costs vary with average transaction size, chargeback rate, and any negotiated terms.

One of the originals

Dwolla was founded in Des Moines, Iowa in 2008 by Ben Milne, and its BBB file independently records the business as started on 17 June 2008. It spent its early years trying to be a consumer alternative to card networks, and then did the sensible thing: it stopped, and turned the plumbing into a product. What it sells today is an API for moving US dollars between bank accounts, and it has been selling that for longer than most of its competitors have existed.

The proposition is narrow and clear. Dwolla does standard ACH, Same Day ACH, and instant payments over the RTP network and the FedNow Service, behind one integration, so a developer can pick the rail per transaction rather than per vendor. It does not touch cards. It says it helps businesses move "tens of billions of dollars every year".

Why bank rails instead of cards

The arithmetic is the whole argument. Card processing is priced as a percentage of the amount, which is defensible on a $40 basket and absurd on a $9,000 invoice. Bank transfers are priced per transfer. For rent, tuition, B2B settlement, insurance premiums, loan disbursement and marketplace payouts, moving off cards is often the single largest cost reduction available to a platform.

What you give up is immediacy and the card networks' dispute machinery. Standard ACH takes one to three business days; Same Day ACH takes hours if it makes the window; RTP and FedNow settle in seconds, all day, every day — but only where both institutions participate. That last clause is the one people forget. Instant settlement is a property of each counterparty's bank, not a switch Dwolla can flip for your whole user base, so verify coverage before you promise it in a product.

The pricing went away

Dwolla used to publish a rate card. It does not now. The pricing page says pricing is "tailored to your transaction volume, rails, and integration needs — not a one-size-fits-all tier", lists no plans, no per-transaction rates and no platform fee, and routes you to sales.

Software directories have not caught up, and still carry the old numbers: around 0.5% per transfer with a floor near 5 cents and a cap near $5, with bundled plans starting around $250 a month. Those come from a rate card that has been taken down and the company does not stand behind them. They may still be roughly right; they may not be. Either way, do not budget from a figure that only exists on third-party sites.

Because nothing is published, the quote conversation carries all the weight. Price each rail separately — standard ACH, Same Day ACH, RTP and FedNow are not the same product. Ask about the platform or minimum monthly fee, the cost of a returned or failed transfer, whether bank-account verification is billed on its own, whether sandbox use is free, and what happens at renewal.

NMI bought it in May

On 19 May 2026 NMI announced it had acquired Dwolla, its sixth acquisition in recent years, adding account-to-account infrastructure, real-time payments, FedNow and payouts to NMI's embedded payments platform. The combined business is described as processing close to $700 billion in annual volume. Roughly 60 Dwolla employees moved across, more than 400 Dwolla customers joined NMI's ecosystem, and Dwolla's chief executive Dave Glaser became NMI's chief operating officer. Terms were not disclosed.

Dwolla's own message to customers is reassuring as far as it goes: NMI "will continue to support Dwolla's existing customers and partners as it integrates Dwolla's capabilities into the NMI platform", and Dwolla customers gain access to NMI's card acceptance — credit, debit, prepaid, HSA/FSA, in-app, in-store and unattended. There is real upside there for a platform that needs both bank and card rails from one vendor.

What is missing from both announcements is any statement that the Dwolla product continues to be sold under its own name once the integration is done. That is not evidence it will not — companies routinely keep acquired brands — but it is an open question three months into a merger, and open questions are what contracts are for. Ask for the roadmap, ask what happens to your pricing at renewal, and ask whether you can export your funding-source and verification records if you decide to leave.

The record

Dwolla's public record is quiet in the way infrastructure companies' records usually are. BBB gives it an A+ on an unaccredited profile with negligible complaint volume, which for an eighteen-year-old money-movement business is genuinely good. Its Trustpilot profile is claimed and empty, which is what a B2B API company's consumer review page normally looks like — read it as no signal rather than a bad one.

The one blemish is old. In March 2016 the CFPB issued a consent order and a $100,000 civil penalty over Dwolla's marketing claims that its data security exceeded industry standards, finding it had not implemented adequate security policies until at least September 2012 and had no written data security plan until at least October 2013. It was the Bureau's first data-security action and it was brought without any breach having occurred or any consumer harm being shown. Ten years and nothing comparable since is a reasonable place to leave it, but a company that once oversold its security posture is a company worth asking direct security questions of.

How to decide

• If your payments are large-ticket and recurring, price them on bank rails. The saving against card interchange is usually the biggest number in the analysis.
• If you need cards, you need a processor too — Dwolla is a complement, not a substitute.
• If you have no engineer, stop here. There is no no-code path.
• Verify RTP and FedNow coverage against your actual counterparties before designing a product around instant settlement.
• Get the quote broken out by rail, and get the platform fee, return fee and verification fee named separately.
• Ask, in writing, what the Dwolla product roadmap is under NMI, what happens to your pricing at renewal, and whether you can export your funding-source records on exit.

The B- is a good product carrying two open questions. Technically, Dwolla does what it says and has done it for a long time; the rail coverage is real and the operating record is clean. Commercially, a provider that has withdrawn its published pricing and changed owners in the same year is one you should sign shorter terms with and ask harder questions of than you would have a year ago.