Integrating Bank Asia's BA Micro Merchant agent-banking app

The records, and how we reach them

Every cash-out, purchase payment, REB electricity bill and mobile top-up an outlet runs through BA Micro Merchant is written to Bank Asia's agent-banking ledger as a dated, authenticated entry. That ledger is the thing worth integrating — not the app's screens, but the structured transaction stream behind them. Bank Asia describes the app as the way a micro merchant checks an agent-banking account holder's balance, hands out cash like a human ATM, and takes purchase payments, all from a phone or tab.

For most teams the useful first move is a clean backfill: walk an outlet's transaction history one day at a time into your own store, then keep it current with a smaller delta job. We get there by analysing the app's own traffic to Bank Asia's backend under the merchant's authorization, and we wrap that in runnable code rather than a report. The sections below name the surfaces, the route we would take, and what you actually receive.

What the app records, surface by surface

These map to services Bank Asia lists for the micro-merchant programme. Granularity below is what an integrator would observe per call or per entry.

Routes in, and the one we would pick

Authorized protocol analysis of the app's backend traffic

The app talks to Bank Asia's agent-banking backend over a documented request/response flow once a merchant logs in with OTP or biometric. Working under the merchant's authorization, we capture and model that flow — login and session, balance lookup, the cash-out and purchase-payment calls, transaction history with paging. Reach is the full ledger above. Effort is moderate; durability is tied to the app version, so we re-validate the captured contract whenever Bank Asia ships an update. This is the route we would build the integration on.

Bangla QR / NPSB rail for the collection side

Where the merchant accepts customer payments, those settle through the national switch and the interoperable Bangla QR standard rather than the app alone. Reading at the rail level gives you inbound collections and settlement records that complement the agent-banking ledger. We use this to cross-check the collection half of the picture.

Native history export as a backfill aid

The app shows merchants their own transaction history and reports. Where that view can be exported, it is a cheap way to seed a first backfill before the live client takes over. It is a supplement, not the spine.

In practice we lead with the protocol route for the live ledger, seed the backfill from whatever export exists, and reconcile collections against the rail. That combination holds up as the app changes because the one moving part — the captured contract — is the part we keep under test.

A transaction-history pull, sketched

Illustrative shape of the backfill client — exact field names and the session handshake are confirmed during the build against the live app, not assumed here.

POST /agent/v2/txn/history          # Bank Asia agent-banking backend
Authorization: Bearer <session_token>   # issued after OTP/biometric login
Content-Type: application/json

{
  "merchant_id": "<outlet code>",
  "window": { "from": "2026-05-01", "to": "2026-05-31" },
  "page_cursor": null,                # null on first call; echoed back to page
  "channels": ["CASH_OUT","PURCHASE","BILL_PAY","RECHARGE"]
}

--- response ---
{
  "entries": [
    { "ref": "BA26051512345", "channel": "CASH_OUT",
      "amount": "3000.00", "currency": "BDT",
      "account_masked": "******8842", "biller_ref": null,
      "ts": "2026-05-15T09:41:22+06:00", "status": "SETTLED" }
  ],
  "next_cursor": "eyJwIjoyfQ==",      # follow until null to complete the day
  "settlement_batch": "2026-05-15"
}

# Backfill loop: for each date in window, page on next_cursor until null,
# upsert by `ref` so a re-fetched day overwrites rather than duplicates,
# then advance to the next settlement_batch.

What lands in your repo

The deliverable is working code first, with the written artefacts alongside it.

• Runnable clients in Python and Node.js for login/session, balance lookup, and the cash-out, purchase, bill-pay and recharge surfaces — the calls an integrator actually invokes.
• A batch backfill harness that walks a date window, pages the history endpoint, and upserts entries by reference, plus a lighter delta job for new activity.
• A reconciliation pass that lines the ledger up against the day-end settlement batch and the Bangla QR collection side, flagging gaps.
• An automated test suite built on recorded fixtures, so a contract change shows up the moment Bank Asia alters a field.
• A normalized schema mapping the app's fields to your own transaction model, with currency, channel and status enums pinned.
• An OpenAPI description of the modelled endpoints, a protocol and auth-flow report covering the OTP/biometric/token chain, interface documentation, and a short data-retention and consent note.

What we plan around for Bank Asia agent banking

Two things about this app shape the build, and we handle both as part of the work:

• OTP and biometric sessions. Login is gated on a one-time code or a fingerprint, and the resulting session lapses. We replicate that handshake and refresh the token mid-run, so a multi-day backfill does not break when a session expires. Credentials and the consenting merchant or agent are arranged with you at onboarding.
• The Bangla QR transition. Bangladesh Bank made the interoperable Bangla QR mandatory at merchant points, retiring bank-specific QR codes. We map both the older and the Bangla QR record shapes so collections reconcile across the changeover instead of splitting into two incompatible histories.
• Agent ledger versus merchant float. A balance check reads a Bank Asia agent-banking account holder's balance, not the merchant's own wallet. We model that distinction explicitly so a disbursement check never reads the wrong number.
• Day-end settlement timing. Entries batch to a settlement date in BDT. We anchor the sync window to that cycle so reconciliation has a stable boundary.

Authorization, Bangladesh Bank, and consent

Bank Asia is a scheduled bank supervised by Bangladesh Bank, and its payment activity sits under the Payment and Settlement System Act, 2024, which the national parliament passed in July 2024 (per the VDB Loi summary) to bring banks and non-bank providers under one framework. The interoperable pieces — NPSB and the mandatory Bangla QR standard — are Bangladesh Bank instruments, not vendor features. Bangladesh has no comprehensive data-protection statute in force as of this writing; a draft has been under consideration, so we do not lean on it and instead anchor access in explicit authorization.

That means the dependable basis is consent: the merchant or agent authorizes the integration, and each transaction already carries the account holder's own OTP or biometric confirmation. We work only against that authorized access, keep consent and access logs, minimize what we retain to the fields a use case needs, and sign an NDA where the engagement calls for one.

Where teams plug this in

• An accounting platform that needs an outlet's cash-out and purchase-payment ledger synced nightly into its books.
• A microfinance lender scoring an agent's turnover from a clean, backfilled transaction history rather than screenshots.
• A treasury tool reconciling agent float and day-end settlement against the bank's batch.
• A biller or utility aggregator matching REB bill-pay collections to its own receivables.

Screens from the app

Public Play Store screenshots, useful for spotting the surfaces named above. Select to enlarge.

Nearby apps in Bangladesh payments

If a single integration needs to span more than Bank Asia, these are the names that come up in the same market. Each holds its own server-side records a unified layer would normalize.

• bKash — the largest mobile financial service; wallet balances, send-money and merchant payment history.
• Nagad — postal-linked MFS with wallet, payment and disbursement records.
• Rocket — Dutch-Bangla Bank's mobile banking, holding account and transaction data.
• Upay — UCB's wallet, with payment and cash-in/out ledgers.
• Tap — Trust Axiata's MFS, covering wallet transactions and merchant collections.
• iPay — an early PSP offering send-money, recharge and bill payments.
• OK Wallet — ONE Bank's wallet with payment and transfer records.
• mCash — Islami Bank's mobile money, holding transaction history.
• NexusPay — Dutch-Bangla Bank's cardless app spanning bank cards and mobile banking.
• SureCash — a payment network used for education and government disbursements.

How this was put together

Checked in June 2026 against Bank Asia's own micro-merchant page, the Google Play listing for the app, reporting on the Bangla QR mandate, Bangladesh Bank's payment-systems pages, and a legal summary of the 2024 act. Sources read in full:

• Bank Asia — Micro Merchant App services
• BA Micro Merchant on Google Play
• The Business Standard — Bangla QR mandatory at merchant points
• VDB Loi — Payment and Settlement System Act, 2024

OpenFinance Lab · interface assessment, June 2026.

Questions integrators ask

Can you backfill an agent's full transaction history, or only new activity going forward?

Both. The first build does a paged backfill across a date range — cash-out, purchase payment, bill pay and recharge entries — so an agent's prior ledger lands in your store, then a lighter delta job picks up new activity after that. The backfill walks day by day to stay aligned with Bank Asia's day-end settlement batching.

The Bangla QR mandate moved merchants off proprietary QR codes — does that change how collection records come through?

It changes the format, not your access to the records. Bangladesh Bank made the interoperable Bangla QR standard mandatory at merchant points, so collections now route through the national switch rather than a bank-specific code. We map both the legacy and Bangla QR shapes during the build so transactions reconcile cleanly across the cutover.

Are balance checks and cash-out records both reachable, or just one of them?

Both are in scope. The app exposes a balance lookup for Bank Asia agent-banking account holders and a cash withdrawal flow the bank describes as a human ATM, plus purchase payment, REB bill pay and mobile recharge. We model the balance lookup against the account holder's ledger, not the merchant's own float, so disbursement checks read the right number.

The login uses OTP and biometric — how does that work with an automated sync?

We replicate the OTP and biometric session handshake the app uses and refresh the session token before it lapses, so a long backfill does not stall halfway through. The credentials and the consenting merchant or agent are arranged with you during onboarding; nothing in the flow asks you to weaken the app's authentication.

Working with us

A first working client for the balance-check and transaction-history surfaces lands inside one to two weeks. You can take the runnable source plus docs and tests for a fixed fee from $300, paid only after delivery once you are satisfied — or skip the build and call our hosted endpoints instead, paying per call with nothing upfront. Tell us the app and what you need from its data and we will scope it. Start a BA Micro Merchant integration