E-commerce connector

ShopiMind connects to your shop to continuously sync customers, orders, products, carts and catalogs. Two options depending on your platform: an official module ready to install, or a custom integration built on the API.

Compatible platforms via an official module

For the main e-commerce platforms, ShopiMind provides a module that installs from your shop's back-office in a few minutes — no code required.

PrestaShop

Official module

WooCommerce

Official module

Shopify

Official module

Magento 1 (OpenMage)

Official module

Magento 2

Official module

If your shop runs on one of these platforms, installing the module is the fastest path: initial sync and incremental flows are handled automatically, without manually configuring the API.

Build a custom connector

If your shop runs on a platform not covered by the official modules — proprietary stack, in-house CMS, private marketplace, or any exotic platform — you can build your own integration through the ShopiMind REST API.

Customers, not contacts

Your integration pushes customers. ShopiMind contacts (unified profiles) are derived automatically from customers — see Customers and contacts.

Resources to sync

Category	Resources	Why ShopiMind needs them
Reference data	Customers groups, Orders statuses, Orders carriers, Products categories, Products manufacturers	Structural data referenced by the main resources.
Customers	Customers, Customer addresses	E-commerce profiles that derive ShopiMind contacts.
Orders	Orders	Purchase history (RFM, scoring, post-purchase triggers).
Carts	Carts	Read active carts (abandonment recovery). Carts are populated server-side from shop events — the public API exposes read-only access.
Catalog	Products	Products referenced for personalization and recommendations.
Promotions	Vouchers	Discount codes usable in scenarios.

Recommended architecture

Phase 1 — Initial import

For a first deployment, import resources in dependency order:

1. Reference data first: customers groups, orders statuses, orders carriers, products categories, products manufacturers.
2. Catalog: products.
3. Customers: customers, then customer addresses.
4. Transactions: orders, vouchers.

Each bulkCreate* endpoint accepts 50 or 100 records per call depending on the resource — large imports are processed in successive batches on your side.

Phase 2 — Incremental sync

Once the catch-up is done, wire writes to the business events of your shop:

Shop event	ShopiMind endpoint
Customer created / updated	bulkCreateCustomers
Customer address updated	bulkCreateCustomerAddresses
Order placed / status changed	bulkCreateOrders
Product created / updated / disabled	bulkCreateProducts
Voucher created	bulkCreateVouchers

All bulkCreate* endpoints upsert: no need to differentiate between creation and update on your side. The upsert key is the supplied identifier (customer_id, order_id, product_id, etc.).

Patterns to follow

Idempotency

Endpoints are designed to be safely called multiple times with the same payload:

• bulkCreate* upserts on the supplied identifier — a retry after a network timeout doesn't create duplicates.
• bulkDelete* silently ignores already-deleted or unknown identifiers.

Per-item validation

Bulks accept partially: an invalid item doesn't interrupt the batch. The response returns sent_count, rejected_count, and the detail of rejected items in rejected_items (with their original index and validation errors). Replay only the rejected items after correction.

Resume on error

If your connector crashes mid-import, it can resume where it stopped using local timestamps: send all records modified since the last successful sync. The upsert merges them cleanly, without duplication.

Asynchronous processing

A 200 response confirms acceptance, not immediate visibility on subsequent reads. Typical processing latency is a few seconds. Avoid issuing a GET right after a POST to verify the data — it's an anti-pattern that produces false negatives.

Performance & monitoring

• Prefer bulks — one bulkCreateCustomers call with 50 customers consumes far less quota than 50 individual calls.
• Parallelize independent resources (e.g. orders and products side by side). Serializing chunks of the same resource doesn't help — ShopiMind's processing queue handles them in order anyway.
• Quota — 600 requests per minute per API key. A 10 req/s import ramp leaves a comfortable margin.
• Monitor the returned sent_count, the rejected_items (worth logging during onboarding), and HTTP status codes: 5XX = retry with exponential backoff, 4XX = fix the data before retrying.

Going further

• API reference — every endpoint detailed.
• Custom data — extend the ShopiMind model with your own fields and objects.
• ShopiMind integration — get your connector validated for the ShopiMind marketplace.