Sage 50 and SPS Commerce integration

What moves between them

The main flow is SPS Commerce into Sage 50. ml-connector polls SPS Commerce for purchase orders and invoices from retail trading partners on a schedule you set, typically 5-15 minutes. Purchase orders from partners translate into Sage 50 purchase orders with the correct vendor and line items, while invoices from you to the retailers post into Sage 50 as vendor payments or expenses once validated. Outbound documents such as order acknowledgments and advance ship notices are built from Sage 50 shipment and delivery data and pushed to SPS Commerce immediately when the sale or shipment is confirmed. Reference data such as vendor codes and item SKUs are aligned in both directions to ensure trading partner document routing matches Sage 50 accounting records.

How ml-connector handles it

ml-connector maintains a Windows process with access to Sage 50's local data files and stores the Windows username and password encrypted. On the SPS side it stores the OAuth client ID and secret encrypted and refreshes the JWT bearer token when a call returns 401. Because Sage 50 is pull-only and desktop-local, ml-connector polls SPS Commerce on your chosen schedule and translates each EDI purchase order, invoice, and advance ship notice into the corresponding Sage 50 transaction, mapping trading partner IDs to Sage 50 vendor codes and item SKUs to Sage 50 inventory items. SPS Commerce rate limits are not public, so ml-connector implements exponential backoff with jitter on 429 responses. Purchase orders and invoices are deduplicated at the application layer using the EDI document number, and every record carries a full audit trail so a failed downstream validation can be replayed once the root cause is fixed.

A real-world example

A mid-sized US supplier sells consumer goods to three major retailers through SPS Commerce. Before the integration, the order desk received EDI purchase orders via SFTP file drop at SPS, manually keyed each line item into Sage 50, and then sent back an order acknowledgment via the same file process. Invoices were prepared in Sage 50 and exported as EDI files for SPS to wrap and route to the retailer. This process took 2-3 hours per day across three trading partners and created data-entry errors that led to payment delays and chargebacks. With Sage 50 and SPS Commerce connected, each purchase order arrives automatically as a Sage 50 PO, inventory is reserved immediately, and invoices are translated and routed to the retailer as soon as the shipment is confirmed. The order desk now focuses on fulfillment rather than data re-keying, and payment disputes have dropped significantly.

What you can do

• •Poll SPS Commerce for purchase orders from your retail trading partners and create matching Sage 50 purchase orders with the correct vendors and line items.
• •Receive invoices and advance ship notices from SPS Commerce and translate them into Sage 50 accounting transactions on the correct accounts and cost centers.
• •Send order acknowledgments and advance ship notices from Sage 50 shipments back to SPS Commerce and your trading partners.
• •Authenticate Sage 50 with Windows credentials and local SDK access, and SPS Commerce with OAuth 2.0 client credentials, on separate polling schedules.
• •Handle SPS Commerce rate limits with exponential backoff, deduplicate by EDI document number, and maintain a full audit trail on every transaction.

Questions

What records move between Sage 50 and SPS Commerce?

Purchase orders, invoices, and advance ship notices flow from SPS Commerce into Sage 50, where they are translated into POs, invoices, and shipment records matched to Sage 50 vendors and items. Order acknowledgments and ASNs are built from Sage 50 and sent back to SPS Commerce for routing to your retail trading partners. The direction and cadence are configurable per document type.

Does ml-connector need to run on the same Windows machine as Sage 50?

Yes. Because Sage 50 is a desktop-only application with no remote REST API, ml-connector's Windows process must run on the same machine or LAN where the Sage 50 company data files live, and it must have exclusive access (the integration user cannot be logged into Sage 50 interactively at the same time).

How does ml-connector handle the differences between Sage 50's local authentication and SPS Commerce's OAuth?

ml-connector stores both credential sets encrypted: the Windows username and password for Sage 50 SDK access, and the SPS OAuth client ID and secret for API calls. It refreshes the SPS JWT bearer token when a call returns 401, and it manages token expiry so outages are avoided. Both credential sets are stored in the cell's encrypted Postgres database.