Selling on Flipkart alongside Business Central usually means two systems that never agree. Orders arrive in the Flipkart seller panel and get keyed into Business Central by hand, so sales orders lag behind approval and pack-by dates. Stock moves in Business Central when a warehouse shipment posts, but the Flipkart listing keeps advertising the old quantity, and the oversell shows up as a cancellation penalty. Settlement reports arrive as files full of commission, collection fee, fixed fee, shipping fee and tax-on-fee lines that nobody reconciles line by line, so the bank receipt never matches the invoice value and the difference quietly parks in a suspense account. Business Central out of the box has no marketplace concept at all: there is no Flipkart order, no listing entity, no place to hold a marketplace order ID next to a sales order, and no way to explain a payout that is smaller than the invoices it settles.

ECOSIRE builds a Flipkart Marketplace Connector as a Business Central AL extension — delivered as a per-tenant extension for a single SaaS environment, or packaged for AppSource-style distribution and on-premises deployment when your estate needs it. The design keeps base application objects untouched: new tables hold marketplace credentials, channel configuration, staged Flipkart orders, listing links and settlement lines; table and page extensions add the Flipkart fields you actually look at (marketplace order ID, fulfilment model, SKU mapping, listing state) onto Sales Order, Item, Sales Shipment and Customer Ledger Entry; and codeunits carry the integration logic behind clean interfaces so the transport layer can be swapped without touching business rules.

Technically, an HTTP client codeunit handles Flipkart Seller API authentication and token refresh, with typed JSON handling through JsonObject and JsonToken rather than string parsing. Inbound work runs on Job Queue Entries at the cadence your volume needs — orders and cancellations on a short interval, listings and settlements on a longer one — and each run writes to a staging table first, so a bad payload is quarantined for review instead of aborting a batch. Documents are created through standard posting routines, never direct table inserts, so your posting groups, dimensions, tax setup and number series behave exactly as they do for any other sales order. Event subscribers on release, warehouse shipment and posting events push shipment confirmations, tracking numbers and carrier details back to Flipkart. Item availability changes trigger a debounced stock and price push, calculated from the availability basis you choose (inventory, projected available, or a reservation-adjusted quantity per location).

Settlement files and payout responses are parsed into settlement lines, matched to invoices, and turned into applied cash receipt journal lines plus separate expense lines for each fee type, so the bank reconciliation clears and fee analysis lives in real G/L accounts instead of a spreadsheet. A dedicated permission set, an integration log with request/response correlation IDs, and retry with exponential backoff round out the operational surface. Where you want the data in Power BI or a Power Automate flow, we expose it through API pages (REST API v2.0 / OData v4) for Dataverse and Power Platform connectivity, and we target the Business Central release wave you are actually running.

Because this is build-to-order, nothing here is a shrink-wrapped download. You request a quotation, we run a scoping call, and we write a specification against your Business Central version and release wave, your fulfilment model (Flipkart-fulfilled versus self-ship), your SKU and category structure, your tax setup, and which fee lines post where. Typical delivery is two to four weeks from confirmed scope. We build on a sandbox environment, hand you the extension for user acceptance testing against your own Flipkart seller account, then install into production with a documented rollback. You receive the AL source code for your version in a git repository you own — no black box, no lock-in — along with technical documentation, a user guide, a training session for the operations team, and a post-go-live support window while the first real settlement cycle runs through the system.