May 29, 2026

Visa Direct Explained: How Push-to-Card Powers Modern Disbursements

Dalia

Head of Growth

Most disbursement debates collapse into “ACH or wire.” That stopped being the right frame around 2018, when Visa Direct quietly turned every debit card in your wallet into a real-time receive address. Today, push-to-card is the rail that powers Uber driver payouts, insurance claim settlements, gig-platform earnings, and the “instant deposit” button in every neobank app.

This is the operator-grade explanation of what Visa Direct actually is, how the rail works under the hood, where it wins against ACH and RTP, where it breaks, and how to decide whether to integrate it directly or through an aggregator. It is not a sales pitch — when push-to-card is the wrong rail for your program, this post will tell you so.

If you’ve already decided push-to-card is the right rail and now you’re comparing platforms, jump to the Prepaid Card APIs vs Push-to-Card vs ACH decision framework. This post is the upstream explainer.

What Visa Direct actually is

Visa Direct is Visa’s real-time push-payments platform. In one sentence: it lets a sender push funds to anyone with a Visa debit card, prepaid card, or eligible credit card, using the same card-network rails that normally move money the other direction (consumer → merchant).

The category name is push-to-card — to distinguish it from the regular “pull” model, where a merchant initiates a charge against a card the cardholder presented. With push-to-card, the recipient is anchored by the card; the sender initiates a payout, and the funds land on the card in seconds to minutes.

Mastercard’s equivalent is called Mastercard Send. American Express has a smaller program called AmEx Send. Functionally they are the same category — the rest of this post focuses on Visa Direct because it has the largest US and global footprint, but everything below applies to Mastercard Send in 95%+ of cases.

Why it exists

Card networks built push rails because four categories of payment had no good answer:

Gig-economy earnings — Uber drivers, Instacart shoppers, DoorDash couriers needed faster than ACH and cheaper than wire.
Insurance claims — fast claim payout is a measurable retention lever, and a 3-day ACH wait kills NPS.
Marketplace payouts — Etsy sellers, Airbnb hosts, eBay sellers expected the money to arrive close to the sale.
Person-to-person transfers — Venmo, Cash App, Zelle all use push-to-card under the hood in some flows.

ACH was too slow. Wire was too expensive. RTP and FedNow weren’t ubiquitous yet. The card networks already had the rails — they just had to run them backwards.

How push-to-card actually works

The transaction flow is similar to a refund, with three structural differences: the sender is not the original merchant, the amount is decoupled from any prior purchase, and the funds are usually available to the recipient much faster than a refund posts.

The five steps

Sender initiates the payout. Your platform (or your aggregator’s platform) sends an Original Credit Transaction (OCT) message to Visa Direct, containing the recipient card PAN, amount, currency, and metadata.
Visa Direct routes the OCT. Visa identifies the issuing bank (the bank that issued the recipient’s card) and forwards the credit message over VisaNet.
Issuing bank posts the funds. Most issuers post the credit to the cardholder’s available balance within seconds to 30 minutes. A small percentage of issuers batch credits and post within hours.
Funds available to the cardholder. The recipient sees the money in their card’s available balance — they can spend it immediately at any merchant, or withdraw it at an ATM if it’s a debit card linked to a bank account.
Settlement happens behind the scenes. Funds flow from the sender’s funding account through Visa’s settlement bank to the issuing bank over the following 1–2 business days. The recipient doesn’t see this; they already have the money.

The critical point: step 4 happens fast. Step 5 happens slow. From the recipient’s perspective, push-to-card is instant. From your finance team’s perspective, it’s still a 1–2 day settlement on the back end. Build your reconciliation around the settlement timestamp, not the recipient-funds-available timestamp.

What you need to send a push-to-card payment

The recipient’s full card PAN (16 digits for most cards). You do not need the CVV, expiration date, or billing address for most push-to-card programs, but some issuers require the expiration for additional verification.
A funding account at your platform or aggregator (USD wallet, bank-debit ACH, or wire-funded).
Some form of recipient identification — name and partial address are common KYC requirements depending on amount thresholds and your platform’s compliance posture.
An aggregator or direct Visa Direct integration. Almost no companies integrate directly with Visa Direct because the certification process, the bank-sponsor requirement, and the ongoing compliance burden make it impractical below high seven-figure annual volumes. The rest of this post assumes you’ll use an aggregator.

When push-to-card wins

Push-to-card is the right rail when:

Recipient has a usable debit card but no relationship with you that justifies collecting their full bank details. Gig workers, contractors, claimants, panelists.
Speed materially changes recipient behavior. Faster claim payout improves NPS. Faster gig payout reduces churn. Faster referral payout lifts completion rate. If “instant” is just a vanity metric in your program, the 2–4% you’re paying for it is dead money.
You can tolerate a 5–18% decline rate — or your aggregator can give you structured decline classification so you can route around it.
Per-recipient value sits in the $20–$2,000 range. Below $20, fixed fees eat the economics. Above $2,000, ACH or wire usually wins on percentage cost.

The honest test: is the lift from “instant” worth the percentage cost over ACH? Run an A/B in your actual program before you decide. The answer is yes more often than ACH defenders admit, and no more often than push-to-card sellers will tell you.

When push-to-card loses

Programs that pay employees on regular payroll. Use payroll. Payroll-tax compliance, W-2/W-4 handling, and benefits coordination are not problems you want to solve on top of a card rail.
International payouts where the recipient is unbanked or carries only a domestic-rail card. Visa Direct cross-border works in the corridors Visa supports, which is roughly 60 destination markets in 2026 — but the recipient needs an enrolled card. Stablecoin or prepaid card APIs usually win when the recipient is genuinely unbanked.
Pure cash-equivalent requirements. Some procurement teams have policies that require ACH or wire and won’t accept card-rail disbursements. Respect the policy; don’t argue.
High-ticket B2B. A $50,000 vendor payment goes by ACH or wire, every time.

The decline-rate reality

This is the section most “what is Visa Direct” posts skip and engineers wish they hadn’t.

In production, push-to-card decline rates of 5–18% are normal, not exceptional. The decline is what makes push-to-card hard to operate — not the integration, not the cost, not the reconciliation.

The categorical decline reasons you should plan for

Prepaid debit card not eligible. Some prepaid programs accept inbound OCTs; many don’t. Aggregators rarely surface this clearly before you try.
Expired card. Recipient hasn’t updated their card on file. ~3–5% of any large recipient list will hit this.
OFAC / sanctions block. The recipient or their card is flagged. The transaction fails silently from your perspective unless your aggregator surfaces the reason.
Issuer-side refusal. Some smaller US credit unions and international issuers refuse inbound OCTs by default. They’re getting better, but the long tail is real.
Daily / monthly velocity limit. The cardholder has already received the max push payouts the issuer permits this period.
Network refusal. Catch-all for issuer responses that don’t fit a standard reason code.

If your aggregator surfaces decline reasons as a single boolean (success: false), your retry logic will be guessing. Demand structured decline reason codes before you sign — and if you’re evaluating aggregators, our deeper write-up walks through what good looks like.

The retry pattern that actually works

The wrong retry pattern: poll the same card on a fixed interval, hoping the issuer changes its mind. They won’t.

The right retry pattern, in priority order:

If the decline is recoverable (velocity limit, temporary issuer flag), retry after the issuer’s window resets — typically 24 hours.
If the decline is structural (prepaid not eligible, expired card, OFAC), stop retrying that card. Either fall back to a different rail (prepaid card API, gift card, ACH) or contact the recipient for an updated card.
Surface the decline reason to your recipient experience. “Your card was declined for X — please add a new card” converts vastly better than a silent retry queue.

A program that runs 10,000 push-to-card payouts a month will see roughly 500–1,800 declines. If you build the retry/fallback path well, your effective delivery rate climbs into the high 90s. If you don’t, every decline is a customer-experience problem and a recipient who blames you, not the card.

Visa Direct vs Mastercard Send — does the difference matter?

For most operators: barely.

The two networks have similar reach, similar pricing, similar decline-rate profiles, and similar settlement timing. Most aggregators route across both networks transparently — you tell them “push to this card” and they figure out which network owns the BIN.

Where the difference matters:

Visa Direct has slightly larger US debit reach (more issuers enrolled by default).
Mastercard Send has stronger reach in parts of Europe and Asia-Pacific.
Cross-border push-to-card corridor coverage differs by network — verify with your aggregator for your specific recipient geographies.

Don’t pick an aggregator based on which network they prefer. Pick based on decline classification, settlement reporting, and the multi-rail story (push-to-card + prepaid card API + ACH through one integration).

What does push-to-card actually cost

Headline interchange-style pricing for push-to-card runs 1.0%–2.5% of the transaction amount, plus a small fixed fee ($0.10–$0.50). That’s the all-in number aggregators quote you — it includes Visa’s network fee, the sponsor bank’s fee, and the aggregator’s margin.

The real cost structure has four layers:

Network fee (Visa, Mastercard) — typically $0.05–$0.15 per OCT, depending on volume tier.
Sponsor bank fee — the bank that holds Visa Direct certification and lets your aggregator originate transactions. Usually $0.10–$0.25 per OCT.
Aggregator margin — what your platform charges on top. Varies widely.
FX spread (cross-border only) — 1–3% above mid-market on the recipient’s currency.

For a $50 referral payout, you’ll see something like $1.00–$1.50 all-in (2–3%). For a $500 contractor payout, $5–$12 all-in. For a $1 microtransaction, the fixed-fee component dominates and push-to-card stops making sense.

The hidden cost most teams miss

Every declined transaction costs operations time, even if the aggregator doesn’t charge you for the decline itself. If your decline rate is 10% and each decline takes 15 minutes of CX/ops time to triage (notify recipient, request new card, re-attempt), the opportunity cost on a 10,000-payout/month program is 250 hours. That’s a full-time headcount you’re paying in operational drag for not building the retry/fallback path properly.

Should you build directly with Visa Direct or use an aggregator?

Direct integration with Visa Direct requires: a sponsor bank, Visa certification (months long), card-data PCI compliance at network-acceptable scope, network membership fees, and an ongoing engineering commitment to maintain the integration as Visa updates the protocol. The threshold where this makes economic sense is roughly $500K+ annual push-to-card volume and a team that already operates payment infrastructure at scale.

Below that threshold, use an aggregator. The good ones (and there are several) abstract Visa Direct, Mastercard Send, ACH, RTP, and prepaid card APIs behind one API with one set of webhooks and one settlement statement. The bad ones charge you aggregator margins on top of a single rail and call it a platform.

GIFQ is in the aggregator category by design. Push-to-card is one of several rails we expose; the payouts API covers gift card, push-to-card, and prepaid card APIs through a single integration with unified webhooks and a single reconciliation surface. If you’re shopping aggregators, our pricing page shows the math at the volume tiers we serve. If we’re not the right answer for your program, the decision framework pillar tells you which rail and which category of platform is.

Where push-to-card is going in 2026

Three trends worth tracking:

Cross-border push-to-card is maturing. Visa Direct now reaches around 60 destination markets directly, and many of the missing ones are served via partner rails that look identical to operators. Five years ago this was a US-only conversation; today it’s a credible international rail for a meaningful subset of corridors.
Stablecoin payouts are emerging as a complementary, not competing, rail. For unbanked recipients in markets where Visa Direct can’t reach (or where local banking infrastructure makes it impractical), USDC and similar stablecoins are starting to do the same job push-to-card does in mature markets. Expect aggregators to add stablecoin alongside card rails over the next 18 months.
Decline-rate classification is finally getting better. Visa and the major aggregators have spent the last two years standardizing structured decline reasons. By end of 2026, “card declined for unknown reason” should be rare in the major aggregators’ APIs. If yours still surfaces it today, that’s a signal about platform maturity.

What ships next

If you’re sizing a push-to-card program: the next read is the Prepaid Card APIs vs Push-to-Card vs ACH decision framework. It runs the math across all four rails so you don’t lock in push-to-card just because it’s the rail you’ve heard of.

If you’re integrating: GIFQ payouts API · pricing · brand catalog.

If you want the engineering-side companion to this post, the gift card API integration guide covers idempotency, webhooks, and the operational primitives that apply equally to push-to-card and gift card APIs.