Reference🧭 Global overviewsIntermediate⏱ 28 min read

🔲 Interoperable QR codes around the world

EMVCo QRCPS and the national standards (QRIS, Thai QR, QR Ph, DuitNow QR, UPI QR, Pix QR, VietQR, KHQR, TR Karekod, QR BCB), static vs. dynamic codes, cross-border links, and what QR acceptance really costs a merchant

A QR code is not a payment method

A payment QR code is a two-dimensional image whose modules encode a string of characters that a phone can read. A QR code is an initiation channel, separate from the payment instrument that carries out the transaction. It holds a payee address, sometimes an amount, and nothing else. The payer’s rights, the cost, the settlement time, and the fraud exposure all depend on the underlying rail the code points to. When negotiating an acceptance contract, that rail is what you are negotiating, and the code sets none of its terms.

  • A QR code on top of an instant credit transfer. QRIS (Indonesia), Pix QR (Brazil), UPI QR (India), VietQR (Vietnam), Thai QR Payment (Thailand): the payment is an irrevocable credit transfer. No chargebacks, no preauthorization, final credit within seconds.
  • A QR code on top of a card. SGQR aggregates international card schemes, and TR Karekod covers both cards and FAST credit transfers: the payment remains a card transaction, with interchange, scheme rules, and a dispute cycle.
  • A QR code on top of a wallet. Alipay, WeChat Pay, GCash: the payment debits an e-money balance, with counterparty risk on the issuer and payout times set by contract.
  • A QR code on top of mobile money. GIMACPAY in the CEMAC zone (Central Africa), TANQR in Tanzania: the balance is safeguarded e-money, and turning it into bank money requires a cash-out.
🔑
The question to ask first, before any integration
The first question in any integration is which rail settles the QR payment, and it has to be answered before any other feature of acceptance is examined. Five properties follow automatically. The first two are whether the payment is irrevocable and whether the payer has a right to dispute. Next come who bears the fraud risk, how fees are structured, and what time the money reaches the merchant’s account. Two QR codes that look identical, in two neighboring countries, can give different answers to all five questions.

The distinction also explains why QR codes spread in markets where card acceptance had remained limited. QR acceptance needs no terminal. A laminated stand is enough, so the upfront cost of accepting payments drops to zero. In Indonesia, 32.71 million merchants were already enrolled on QRIS in 2024 (Bank Indonesia), a scale no POS terminal fleet reaches in an emerging market. In Bolivia, 88.4% of QR payments are under 500 bolivianos and 49.7% are under 50 bolivianos (Banco Central de Bolivia, Informe de Vigilancia del Sistema de Pagos 2025). There, QR codes handle small tickets that cards could not process at a cost acceptable to merchants.

The same gesture: a QR codethe customer scans and approvesA · The platform owns the loopB · The state owns the standardProprietary QR codereadable by one app onlythe QR codeStandard QR codeany compliant app can read itPlatform walletprepaid balance in the appthe fundsBank accountthe funds stay at the bankInternal settlemententry in the wallet's ledgersettlementSettlement on the public railcentral bank moneyThe merchant signs upfee set by the platformaccessThe merchant is already on boardprice capped by public rulesthe data stays captiveone standard, many appsinteroperability: none by defaultinteroperability: built inThe gesture is the same. The difference is who owns the rail, and therefore who sets the price.On side A, leaving the platform means losing the whole network. On side B, switching apps costs nothing.
⚠️
A QR code can also be a walled garden
Interoperability means that any licensed app can read the code displayed by any merchant in the country. The code’s technical format does not deliver that on its own. For a decade, China ran two QR networks, Alipay and WeChat Pay, that could not read each other’s codes. A merchant had to display two codes and sign two contracts. It took a PBoC mandate, implemented bilaterally by Ant, Tencent, and China UnionPay from 2021, to make their offline barcodes scannable by each other’s apps. “QR” and “interoperable” are two independent properties, and this guide covers only cases where interoperability has been mandated.

EMV QRCPS: the standard beneath the standards

The EMV QRCPS specification defines the structure of the data encoded in a payment QR code, and the two ways the code can be presented. Almost every national QR standard in this guide claims EMVCo compliance, including QRIS, Thai QR Payment, QR Ph, DuitNow QR, LANKAQR, KHQR, and Pix QR. The reference document is the EMV® QR Code Specification for Payment Systems (EMV QRCPS), published by EMVCo in two separate volumes, both at version 1.1, dated November 27, 2020 (EMVCo, public specifications).

Merchant-Presented Mode (MPM)Consumer-Presented Mode (CPM)
Who displays the codeThe merchantThe customer, on their phone screen
Who scansThe customer, with their appThe merchant, with an optical scanner or camera
Merchant equipmentNone (a sticker) or a screenA 2D scanner or a camera-equipped register
Content encodingText, positional TLV in the clearBase64-encoded BER-TLV; application template with a cryptogram
Type of tokenPayee address, reusableSingle-use, short-lived payment token
Typical use caseMicro-merchants, markets, bills, mobile e-commerceHigh-throughput checkouts, transit, chain retail
Dependence on the customer’s connectionHigh: the customer must be onlineLow: the code can be generated in advance, offline
The two EMV QRCPS modes are not two options of the same product

MPM dominates deployments worldwide because it shifts the equipment cost from the merchant to the customer, who already owns a phone. Its structure is a positional TLV. Each data object is a sequence of identifier (2 digits) + length (2 digits) + value, strung together with no separator. The content can be read in the clear from the image alone, with no key and nothing to decrypt. An MPM QR code holds no secret and is not authenticated. Anyone can copy one exactly or swap in a different one, and nothing in the format prevents it.

Anatomy of an EMVCo Merchant-Presented Mode QR code
000201                      tag 00: Payload Format Indicator, value "01"
010212                      tag 01: Point of Initiation Method
                                     11 = STATIC (reusable)
                                     12 = DYNAMIC (single use)
26XX....                    tags 02 to 51: Merchant Account Information
                                     one template per scheme or per national
                                     standard; this is where the payee's real
                                     address lives (alias, VPA, Pix key,
                                     account number, wallet identifier)
52045812                    tag 52: Merchant Category Code (MCC, ISO 18245)
5303360                     tag 53: transaction currency (ISO 4217 numeric)
540512500                   tag 54: amount (absent on a static QR)
5502011                     tag 55: tip / convenience fee indicator
5802ID                      tag 58: country (ISO 3166-1 alpha-2)
5913WARUNG BU SRI           tag 59: merchant name (25 characters max)
6008SURABAYA               tag 60: merchant city (15 characters max)
62XX....                    tag 62: Additional Data Field Template
                                     01 bill number
                                     03 store label
                                     05 reference label (reconciliation)  <<< KEY
                                     07 terminal label
6304A1B2                    tag 63: CRC-16/CCITT-FALSE over the WHOLE payload,
                                     "6304" included, value excluded
ℹ️
Tag 62-05, the reconciliation key to write into the contract
The Reference Label in the additional data template is the only field that links a credit to the merchant’s account back to an order in its POS or order system. Many national implementations make it optional, and many aggregators do not pass it through in their webhooks. Before signing, check that the field is actually populated, that the rail carries it, and that it comes back in the settlement file. Without it, reconciling a QR payment relies on amount and timestamp alone, and two sales for the same amount in the same minute can no longer be told apart.

CPM is built differently. The customer’s phone generates a Base64-encoded BER-TLV, whose application template carries the application identifier, the account data, and a cryptogram. It is the EMV logic of a contactless card, carried over into an image. The cryptogram authenticates the transaction, which no MPM code can do, and the merchant does the scanning, so it needs a scanner. CPM therefore takes hold where throughput matters and checkout equipment already exists, such as supermarket registers and transit, rather than where acceptance has to stay free to deploy.

⚠️
“EMVCo-compliant” does not mean “interoperable”
Two QR codes that fully comply with EMV QRCPS can still be unreadable to each other’s apps, because the content that matters sits in the merchant account templates (tags 02 to 51), whose allocation and format each national standard sets. EMVCo compliance guarantees that a parser can read the payload structure. It never guarantees that a Thai app has the routing rules tied to an Indonesian identifier. Cross-border interoperability comes from a commercial and clearing agreement, not from the format.

Static and dynamic: two products, not two settings

Tag 01 of the payload, the Point of Initiation Method, takes only two values, which mark two different uses of the same format. A static QR code, value 11, is a printed address with no amount and no reference. It can be reused indefinitely and costs nothing to distribute. A dynamic QR code, value 12, is generated by the register for a single transaction. It carries the amount, the currency, and the reference, and it expires.

DimensionStatic QR (tag 01 = 11)Dynamic QR (tag 01 = 12)
AmountEntered by the customer, a source of errors and underpayment fraudSet by the register, cannot be changed
Order referenceUsually missingCarried in tag 62-05, inserted by the POS system
Accounting reconciliationManual, by amount and timeAutomatic, one to one
Deployment costNext to nothing: printing, a stand, a stickerIntegration with the register or terminal, plus a screen or printer
Physical riskSticker swapping (QR swap)None: the code exists for only a few seconds
Payment confirmationNotification in the merchant app, checked by eyeFeedback to the register: the POS knows the payment went through
Use caseMarket stalls, taxis, tradespeople, tips, donations, billsChain retail, restaurants, e-commerce, transit
How the two modes differ in day-to-day operations
⚠️
Static QR codes give no feedback to the register, and fraudsters exploit that everywhere
With a sticker, there is no technical link between the payment and the register. The merchant accepts the sale on the strength of a screen the customer shows them. Faked screenshots and fake confirmation apps are routine fraud in every QR market, from Jakarta to Lagos. The operational countermeasure is to accept only the notification on the merchant’s own phone, or the audio alert from a soundbox, and never the payer’s screen. The second type of fraud is sticker swapping. A fraudster pastes their own code over the merchant’s and collects the payments instead, and the swap only shows up on the statement.

This split has a design consequence that is often missed. A dynamic QR code is a separate product from a static one, aimed at a different type of merchant, not an upgraded version of the same product. Countries where adoption took off started with static codes, precisely because they cost nothing and need no equipment. Only then did they connect dynamic codes to chain retailers’ registers. The reverse order, dynamic first, is a fairly reliable sign of a failed rollout. It limits acceptance to merchants that already have a register that can be integrated, which is the part of the market cards were already serving.

ℹ️
The physical limit of scanning, and how to get around it
A scan takes a few seconds: the customer has to take out the phone, unlock it, open the app, aim, and confirm. That delay is fine at a market stall and a deal-breaker at a subway turnstile. Bank Indonesia therefore launched QRIS Tap in 2025, an NFC extension of QRIS announced at about 0.3 seconds per transaction (Bank Indonesia). Brazil made the same move with Pix por Aproximação, which triggers a Pix payment when the phone is tapped, with no code to scan. Both responses follow the same pattern. Where QR codes hit a speed limit, the national standard adds contactless without changing the rail. Only the initiation step changes, from scanning to tapping, and settlement still runs over QRIS in one case and Pix in the other.

National QR standards around the world

A national QR standard always fits the same definition: a single code displayed by the merchant, readable by every licensed app in the country. What varies is the settlement rail underneath, the authority that sets the standard, and how strictly it is enforced. The table below covers the live standards in the main markets. Every figure comes with its source and year.

StandardCountryOperatorSinceSettlement rail
QRIS (Quick Response Code Indonesian Standard)IndonesiaBank Indonesia with ASPI2019Credit transfer / e-money, interoperable
Thai QR Payment (Standardised Thai QR Code)ThailandBank of Thailand / National ITMX2018PromptPay (instant)
QR PhPhilippinesBangko Sentral ng Pilipinas with PPMI2019InstaPay (instant)
DuitNow QRMalaysiaPayNet2019DuitNow (instant) + UnionPay acceptance
SGQRSingaporeMAS / IMDA through the Singapore Payments Council2018Multi-scheme: PayNow, NETS, card schemes
UPI QRIndiaNational Payments Corporation of India2016UPI (instant)
Pix QRBrazilBanco Central do Brasil2020Pix / SPI (instant)
VietQRVietnamNAPAS2021NAPAS 247 (instant), address = bank account
KHQRCambodiaNational Bank of Cambodia, on Bakong2020Bakong (dual-currency: riel / US dollar)
LANKAQRSri LankaCentral Bank of Sri Lanka / LankaClear2018Domestic interbank
MMQRMyanmarCentral Bank of Myanmar2025Switch linking 11 wallets
tarusQRBruneindpx, under the Brunei Darussalam Central Bank2025Interbank and wallets
JPQRJapanPayments Japan Association, backed by METI2019Multi-provider, one code per merchant
Taiwan Pay (台灣Pay)TaiwanFinancial Information Service Co.2017Direct debit from the bank account
TR KarekodTurkeyBKM2020Card and FAST instant credit transfer
QR PlatbaCzechiaČeská bankovní asociace2012Credit transfer order
IPS pokaži / IPS skenirajSerbiaNarodna banka Srbije2020IPS NBS (instant)
qvikHungaryMagyar Nemzeti Bank / GIRO Zrt.2024AFR (instant)
RoPayRomaniaTRANSFOND S.A.2025Plăți instant
Transferencias 3.0 / PCTArgentinaBanco Central de la República Argentina, COELSA clearing2020-2021Credit transfer (not card)
QR SimpleBoliviaBanco Central de Bolivia with ASFI2019Interbank, including cooperatives and credit unions
CoDiMexicoBanco de México2019SPEI (instant)
GhQRGhanaGhIPSS2020Interbank and mobile money
TANQRTanzaniaBank of Tanzania2022TIPS (instant)
GIMACPAYCEMAC zone (6 countries)GIMAC, under the aegis of the BEAC2020Card, mobile money, and transfers on a single rail
Bangla QRBangladeshBangladesh Bank, on the NPSB platform2020Domestic interbank
FonepayNepalFonepay Payment Service Ltd (F1Soft group)2020Private network turned de facto standard
ELQRKyrgyzstanNational Bank of the Kyrgyz Republic2022National instant payment rail
Fiji QR Code SchemeFijiReserve Bank of Fiji–Domestic interbank
QMP (Qatar Mobile Payment)QatarQatar Central Bank–Interoperable instant mobile switch
聚易用 Simple PayMacaoAutoridade Monetária de Macau–Aggregates local QR payment instruments
Interoperable national QR standards (operator, year, underlying rail)
32.71M
merchants enrolled on QRIS, with 50.50M users
Bank Indonesia, 2024 data
891M
QR payments in Bolivia in 2025 (+131% year over year), or about 28 per second
Banco Central de Bolivia, Informe de Vigilancia del Sistema de Pagos 2025
~90M
accounts scanning VietQR codes (October 2025), volume up 52% year over year
NAPAS, 2025
> 9 in 10
Fonepay’s share of QR payments in Nepal
Nepal Rastra Bank, as reported by the Nepali press
> XAF 600B
GIMACPAY transactions in 2025; 124 participants, 37M connected wallets
GIMAC / BEAC, 2025–2026
~9,900/day
average daily CoDi transactions in Mexico in 2024, an admitted adoption failure
Industry sources, 2025; Banxico and the ABM (Mexico’s bankers association) publicly acknowledge low awareness (Nov. 2025)
ℹ️
Saudi Arabia: a national standard that is not EMVCo-based
Saudi Payments, a subsidiary of the central bank (SAMA), picked the software vendor HPS on February 27, 2020, to build the unified national QR platform linking banks, wallets, and fintechs. The platform is described as running on ISO 20022 messaging rather than EMV QRCPS, on a request-to-pay model. The merchant presents a request, and the customer reviews it, then accepts or rejects it. That architecture sets Saudi Arabia apart from almost every other standard in this guide. The platform handles both static and dynamic codes, and the stated principle is the same as everywhere else: one code, whatever the customer’s provider or scheme (HPS, press release, February 27, 2020).
🔑
The Czech Republic standardized payment QR codes in 2012
QR Platba, the Czech Banking Association’s standard, encodes a credit transfer order, not a card transaction. It predates the EMVCo specifications by half a decade, and every banking app in the country reads it today, on bills as well as at the register. That head start undercuts the usual view of the payment QR code as an Asian or EMVCo invention. Central Europe had a QR payment standard before the markets usually credited with the idea.

Who sets the standard: three governance models

The governance of a national QR standard comes down to who publishes the specification and decides whether it is mandatory. No national QR standard grew out of the market on its own. In every country, an authority made the call, and who that authority is predicts the adoption rate fairly well, along with fee levels and how stable the system will be over ten years. Three setups cover most cases.

🏛️
Central bank mandate
QRIS, Pix QR, Transferencias 3.0, CoDi, TANQR, IPS pokaži, qvik. The central bank publishes the standard and makes participation mandatory. This is the fastest model, and the only one that lets a regulator set merchant pricing. Bank Indonesia published QRIS on August 17, 2019, and made it mandatory as of December 31, 2019.
🏦
Interbank consortium
TR Karekod (BKM), QR Platba (ČBA), SGQR (Singapore Payments Council), JPQR (Payments Japan Association), Taiwan Pay (FISC), QR Ph (BSP with PPMI). An industry body owns the standard. It is slower to get going, but banks accept it more readily, because they write the standard instead of having it imposed on them.
🚀
Private de facto standard
Fonepay in Nepal is the textbook case. More than 9 in 10 QR payments run over a private network rather than public infrastructure. The model works but is fragile, because the regulator always ends up wanting to take back control, as Kazakhstan did with Kaspi.
⚖️
Interoperability forced after the fact
Argentina and China. The regulator did not build a rail. It forced dominant incumbents to make their QR codes readable by each other: Mercado Pago through Transferencias 3.0 (BCRA), and Alipay, WeChat Pay, and UnionPay through a PBoC mandate. The hardest of the four, politically and technically.
2012
QR Platba (Czech Republic)
The Czech Banking Association standardizes a QR code that encodes a credit transfer order, five years before the EMVCo specifications.
Sept. 2016
台灣Pay QR Code 共通支付 standard (Taiwan)
FISC publishes the common standard; the 台灣Pay brand follows on November 16, 2017.
2018
Thai QR Payment and SGQR
Thailand builds its standard on PromptPay; Singapore publishes the world’s first unified QR code combining several domestic and international schemes on a single label.
2019
The year of mandates
QRIS (Bank Indonesia, published August 17, mandatory from December 31), QR Ph (BSP), DuitNow QR (PayNet), CoDi (Banxico).
Feb. 27, 2020
Saudi national QR platform
Saudi Payments picks HPS to build the unified platform linking banks, wallets, and fintechs (ISO 20022, static and dynamic).
2020
Pix (Brazil) and KHQR (Cambodia)
Banco Central do Brasil mandates the EMVCo QR code as part of core Pix; the National Bank of Cambodia launches KHQR on Bakong.
Nov. 27, 2020
EMV QRCPS v1.1
EMVCo publishes version 1.1 of both volumes, Merchant-Presented Mode and Consumer-Presented Mode.
2021
VietQR, and the end of China’s walled gardens
NAPAS launches VietQR; in China, the PBoC mandate makes the offline barcodes of Alipay, WeChat Pay, and UnionPay scannable across all three apps.
2022
Regional Payment Connectivity (ASEAN)
Five Southeast Asian central banks sign a memorandum of understanding covering the bilateral QR links already in progress.
Sept. 1, 2024
qvik (Hungary)
The central bank requires the acceptance layer to be available in every banking app, with no separate app and no fees for either merchant or customer.
March 26, 2025
Nexus Global Payments
A governance entity is incorporated in Singapore to replace the web of bilateral links with a multilateral model.
July 5, 2025
JPQR Global
Launched at the Expo 2025 Osaka, Kansai site, it connects JPQR with KHQR, then with QRIS from August 17, 2025.
2025
The latest wave
MMQR (Myanmar), tarusQR (Brunei), RoPay (Romania), QRIS Tap (Indonesia), and a pilot of Kazakhstan’s unified interbank QR system.
⚠️
A regulatory mandate does not create usage
GhQR (GhIPSS, 2020) is technically live, but commercial use has been disappointing. Bangla QR (Bangladesh Bank, on NPSB, 2020) recorded 77,165 transactions worth Tk 22.02 crore on June 30 and July 1, 2026, right at the mandatory migration deadline for proprietary QR codes. That volume is negligible for a country of Bangladesh’s size (Bangladesh Bank / BSS). CoDi is free, interoperable, and built on SPEI, yet it handles only about 9,900 transactions a day. TANQR has gone the other way, because joining TIPS was itself mandatory. The four cases point to the same conclusion: adoption comes from mandating the settlement rail, not the code on display.

What QR acceptance really costs

The cost of accepting a QR payment is the fee charged to the merchant, the merchant discount rate (MDR), plus the operating costs that come with the channel. QR is often sold as the free channel, but the actual rate cards say otherwise. How the fee is set varies widely by country, under five regimes. Indonesia has a regulated, published rate, India a legal ban on any fee, and Saudi Arabia a regulatory cap on the underlying card rail. Hungary chose zero fees mandated by the central bank, while Brazil and most other markets leave the price to be negotiated freely with the provider.

Merchant categoryQRIS MDR
Micro business (UMI), transaction ≤ Rp500,0000 %
Micro business (UMI), above Rp500,0000,3 %
Small, medium, and large businesses0,7 %
Education0,6 %
Public retail fuel stations0,4 %
Public services, government agencies, social programs0 %
QRIS merchant pricing, the market’s only complete public rate card (Bank Indonesia)
  • Indonesia: the rate card above, with a cap of Rp10 million per transaction. Providers can add their own daily or monthly limits based on their risk assessment (Bank Indonesia).
  • India: fees are prohibited by law. Since January 1, 2020, Section 10A of the Payment and Settlement Systems Act, 2007 and Section 269SU of the Income-tax Act, 1961 have barred any charge to either payer or payee on UPI and RuPay debit payments. The direct result: on UPI, acceptance is not a profit center, and the business model shifts to lending, product distribution, and data.
  • Saudi Arabia: regulation caps the merchant fee on the mada card rail at 0.80%, with a ceiling of about SAR 40 per transaction (SAMA). That cap, not the QR channel, drives the economics of local acceptance.
  • Hungary: qvik charges no fees to either merchant or customer, by regulatory design (Magyar Nemzeti Bank). It is a direct and openly declared attack on card interchange.
  • Mexico: CoDi is completely free on both sides. It still has not taken off, because zero fees make up neither for its absence from banking apps nor for the lack of any incentive for merchants.
  • Brazil: Pix is free for consumers by decision of Banco Central do Brasil, but on the merchant side, each participating institution sets its own price, not the central bank. Pricing varies more between Brazilian providers than between countries, so the comparison that matters happens within the market.
🔑
The real cost of QR is almost never the fee
On an A2A rail, where money moves from account to account for little or no fee, the costs shift to operations. Reconciliation becomes manual without tag 62-05. Refunds are outgoing credit transfers, with AML/CFT checks, an approval step, and their own risk of paying the wrong person. Treasury has to handle instant credits into an account that then needs sweeping. Training cashiers to check the notification adds to the bill. In a multi-country rollout, integration becomes the biggest cost, because the work has to be redone for each national standard.
⚠️
No chargebacks, so no recourse for the customer
When a QR code runs on an instant credit transfer, the payment is irrevocable. The merchant therefore faces no chargeback risk, holds no reserve, and tracks no dispute ratio. The flip side is customer service. A customer who disputes a payment has no rights to invoke, so the only place to complain is the merchant. Handling disputes then depends on a clear refund policy, a stated response time, and a dedicated channel. Without them, the complaint goes nowhere. On these rails, the merchant’s own policy is the payer’s only means of recourse.

Cross-border links: the bilateral web and what comes next

Cross-border QR acceptance lets a traveler pay a foreign merchant with their usual payment app. Today it relies mostly on a web of bilateral agreements between central banks, not on a global standard. The traveler pays by scanning the host country’s national QR code with their usual domestic app. The merchant keeps its code, contract, and equipment. It receives a domestic credit, in its own currency, into its usual account.

An ASEAN cross-border QR payment, end to end
Merchant
Displays its usual national QR code
An Indonesian QRIS code, for example: no extra label, no extra contract, no special equipment
Traveler
Scans with their domestic app
The app reads the EMVCo payload, identifies the country from tag 58, and detects that the payee is not domestic
Payer country’s switch
Routes the transaction to the bilateral corridor
NITMX in Thailand, PayNet in Malaysia, NAPAS in Vietnam, BCS in Singapore. The national operator acts as the gateway
Currency conversion
Calculates the rate and displays it before confirmation
This is where the real cost of cross-border payments lies: the rate applied, and how transparent it is, varies from one link to the next
Payee country’s switch
Injects the payment into the domestic rail
From the merchant acquirer’s point of view, the payment is once again an ordinary local transaction
Settlement banks
Settle in local currencies
Interbank settlement runs through designated settlement banks in local currencies, which reduces reliance on the US dollar
Merchant
Gets credited as for a domestic sale
Same timing, same file format, same account: on the collection side, the cross-border payment is invisible

The documented web of links is dense. Live links include Thailand–Cambodia (2020), Thailand–Vietnam (2021), Thailand–Malaysia (2021), Thailand–Singapore (2021), and Thailand–Indonesia (2022). They also include Malaysia–Singapore (2023), Malaysia–Indonesia (2023), Cambodia–Laos (2023), Cambodia–Vietnam (2023), Indonesia–Singapore (2023), Malaysia–Cambodia (2024), and Laos–Thailand (2024). In all, 29 QR and P2P links within ASEAN or with outside partners were counted as of December 2025 (regional sources, 2025–2026). Cambodia, with KHQR, has the densest network for its size, with links to Thailand, Laos, Vietnam, Malaysia, China, Singapore, and India.

  • Outside ASEAN, India exports its QR code. UPI acceptance is live in Bhutan (2021, the first country in the world to accept UPI for merchant payments), the UAE through NeoPay (2022), Singapore, Nepal, Sri Lanka, Mauritius (2024), and Qatar through QNB (2024). The Nepal corridor, opened in February 2024, carried 134,701 transactions worth Rs 321 million in five months (Fonepay / NPCI International).
  • Japan came late but moved fast. JPQR Global, launched on July 5, 2025, at the Expo 2025 Osaka, Kansai site, links JPQR with KHQR, then with QRIS from August 17, 2025.
  • Laos built the UnionPay specifications into its national standard in late 2024 to capture Chinese tourist spending. It is the classic case of a small market opening up out of economic necessity.
  • A competing private route exists: Alipay+ (Ant International, since 2020). It claims more than 2 billion accounts reachable through some 50 partner wallets, more than 150 million merchants, and more than 220 markets covered (alipayplus.com, checked in 2026). For a merchant, that means one integration instead of N bilateral agreements, at the cost of depending on a private intermediary.
  • The Philippines and Myanmar remain outside the ASEAN QR network, with immediate operational consequences for anyone doing business in those markets.
⚠️
FX is the unsolved problem of cross-border QR
Routing the payment works, and the merchant gets paid in its own currency. But the exchange rate applied, and how transparent it is, varies from link to link, and no common regional framework standardizes it yet. Launch announcements never mention this, and finance teams find out at their first reconciliation. On any meaningful cross-border volume, the only way to measure the cost is for the acquirer to disclose the FX margin line by line. A monthly average rate blends links with different margins and hides the gap.
ℹ️
Nexus: the planned shift from bilateral to multilateral
The bilateral model requires n×n agreements, whose cost, lead time, and fragility grow with every country added. Nexus, designed by the BIS Innovation Hub (blueprint completed in July 2024), proposes a hub-and-spoke model in which a single connection reaches every member instant payment system. Nexus Global Payments was incorporated in Singapore on March 26, 2025, to run its governance. Its founding members are India, Malaysia, the Philippines, Singapore, and Thailand, later joined by Bank Indonesia. The system is not yet live, and legal and technical milestones are still pending. No acceptance project can rely on Nexus until they are met. In 2026, bilateral links are the only real cross-border coverage.

What breaks in production

Operational incidents in QR acceptance are failures that keep a payment from going through or keep the funds from being identified. They recur with great regularity from one country to the next. They have less to do with cryptography than with the physical and accounting setup around the code. The table below lists them, with the root cause and the fix for each.

SymptomRoot causeWhat to do
No app can read the QR codeCRC (tag 63) miscalculated; the classic mistake is leaving “6304” out of the calculationTest the payload with at least three different issuers’ apps, not just one
The code scans, but the payment is declinedCurrency (tag 53) or country (tag 58) does not match the payee accountCheck that the generator maps currencies to numeric ISO 4217 codes (not alphabetic ones)
Merchant name cut off on the customer’s screenTag 59 is limited to 25 characters, tag 60 to 15Pick a short display name that the customer will recognize, or they will abandon the payment
Half of all payments fail to reconcileTag 62-05 is missing, not carried by the rail, or not returned in the settlement fileMake it a contract requirement and test it end to end, from QR code to statement
The merchant receives the wrong amountStatic QR: the customer enters the amountSwitch to dynamic codes once the average ticket justifies it; until then, check every notification
The merchant receives nothing even though the customer paidA fraudster swapped the sticker (QR swap)Sealed stand or code under glass, a daily visual check, and mandatory daily reconciliation
Fake payments accepted at the registerThe cashier relies on the customer’s screenConfirm only on the notification the merchant receives, or on a soundbox audio alert
Refund sent to the wrong personThe refund is an outgoing credit transfer, keyed in by hand for lack of a reliable aliasOnly allow refunds triggered from the original transaction, never keyed in manually
Common failure modes in QR acceptance
⚠️
Static QR codes never expire, and they stay in circulation
A static QR code is a permanent address. It outlives a store closure, the end of an acquiring contract, a flyer printed three years ago, and an archived PDF invoice. Revoking codes requires an inventory of every code issued, recording the store, issue date, and distribution channel for each. Without that inventory, a business cannot tell where its payment addresses are still in circulation, or which of them pay into an account it no longer uses.
ℹ️
Offline is the blind spot of A2A rails
A static QR code needs the customer to be online, and a dynamic one needs the register to be online. If the network goes down, the channel stops working, with no fallback comparable to card schemes’ stand-in processing. National workarounds exist but are partial. India built UPI 123PAY (IVR, missed call, embedded app, sound-based payments) for feature phones, and UPI Lite X for offline NFC payments under a low limit. Most standards have no answer to this problem, and coverage has to be checked market by market.

One last architectural choice determines long-term dependence. A standard encodes either a bank account or a wallet ID. VietQR encodes a bank account. That choice explains why Vietnam did not become dependent on a dominant wallet, unlike Indonesia or the Philippines. Taiwan Pay and QR Platba made the same choice. When a standard encodes a wallet ID, the ecosystem consolidates around two or three dominant players, and merchants gradually lose the ability to play one provider off against another.

Getting connected: who to talk to and what to demand

For a merchant, connecting to a national QR standard means obtaining an acceptance ID that the country’s licensed apps recognize. In nearly all these markets, direct access to the standard is limited to licensed institutions: banks, payment institutions, and e-money issuers. A merchant or foreign provider cannot get an ID from the central bank. It has to go through a participant. The first job in any project, before any technical question, is therefore to identify which entities are allowed to onboard a merchant in the target country.

🇮🇩
QRIS (Indonesia)
Bank Indonesia publishes the standard, and the Asosiasi Sistem Pembayaran Indonesia (ASPI) runs the ecosystem. Merchants enroll through a licensed provider (a bank or a PJP, Indonesia’s licensed payment service provider). The MDR rate card is public and regulated, and therefore not negotiable.
🇮🇳
UPI QR (India)
NPCI runs the rail, but direct access is limited to banks. A foreign company goes through a TPAP (third-party app provider) partnership or a PSP-and-bank pairing. Merchant fees are prohibited by law, so acquiring generates no revenue.
🇧🇷
Pix QR (Brazil)
Banco Central do Brasil requires any institution with more than 500,000 accounts to participate and makes the EMVCo QR code mandatory. Each participant sets its own merchant pricing, so the negotiation happens with the provider, not the regulator.
🇵🇭
QR Ph (Philippines)
The three-tier structure is unusual. The BSP (the central bank) sets the framework, the industry association PPMI handles governance, and private switches (BancNet) run operations. Each tier deals with different issues, and a request sent to the wrong one can cost several months.
🇲🇾
DuitNow QR (Malaysia)
PayNet runs the standard, and it is mandatory. Banks and wallets must all accept the same code. Four cross-border links (Thailand, Singapore, Indonesia, Cambodia) and UnionPay acceptance come with it.
🌍
GIMACPAY (CEMAC zone)
Instruction BEAC 001/GR/2018, from the regional central bank, mandated interoperability. Cards, mobile money, and transfers converge on a single rail across six countries. The way into the CEMAC zone is through GIMAC, not by negotiating with banks one at a time.
  • Identify the settlement rail before anything else: instant credit transfer, card, e-money, or mobile money. Everything else follows from that.
  • Establish who is allowed to onboard a merchant: get the list of licensed participants published by the central bank or the operator, and check that the prospective partner is still on it.
  • Get the applicable pricing: regulated and public (Indonesia), banned by law (India), capped (Saudi Arabia), or unregulated (most markets). The answer changes the negotiation completely.
  • Test tag 62-05 end to end: QR generation, notification, settlement file, bank statement. Until the reference shows up at the last step, you cannot reconcile.
  • Check how voids and refunds work: on an irrevocable rail, a refund is not a reversal but an outgoing credit transfer, with its own compliance checks and timing.
  • Check the target market’s cross-border status: whether inbound QR payments are accepted, which origin countries are covered, and who bears the FX risk.
  • Keep a register of every static code issued from day one, with store, date, and distribution channel. You cannot rebuild that inventory after the fact.
✅
What successful rollouts have in common
Markets where interoperable QR has taken hold share three features. The first is an instant payment rail that institutions were already required to join, as with TANQR on TIPS, Pix, or UPI. The second is a zero or regulated acceptance cost. The third is distribution through the banking apps customers already use, rather than a separate app to install. GhQR, CoDi, and Bangla QR are cases where one of the three is missing. The standard is published and the codes are on display, but usage has not followed.
🔑
QR is not a stepping stone to something else
QR is sometimes described as a stopgap on the way to contactless. What has happened on the ground does not bear that out. QRIS Tap, Pix por Aproximação, and UPI Lite X add NFC without changing the rail or the standard. What these countries have built to last is an interoperable rail, free or nearly free, addressed by alias. The trigger can be a scan, a tap of the phone, or a link. The initiation channel evolves, while the settlement architecture stays in place.