ShopAssistCatalog sync
How your products reach the assistant, and how fast a change shows up.
We pull, you push nothing. Our server calls your store API with a read-only credential you create. Your store never calls us, and your storefront never passes catalog data through. There is no webhook for you to host and no queue for you to run.
What we read
Public catalog fields only. Title, description, price, availability, images, product URL, options and variants. The same data a shopper sees on the product page.
We do not read customer records, addresses or payment data. On Shopify the catalog comes from the Storefront API, which has no access to any of that.
Orders are separate. We read them with their own read-only scope, and only to report which orders came out of a conversation. Turn that scope off and the assistant still works, without revenue reporting.
How often
There are two modes. We pick the one that fits the catalog, and you can ask us to switch.
| Mode | How it works | When a change shows up |
|---|---|---|
| Direct | We read the catalog while the shopper chats, and cache it for a short time. | Within minutes |
| Indexed | A scheduled job pulls the catalog and builds a search index. | At the next run |
Direct mode suits catalogs up to a few hundred products. Indexed mode is for large catalogs, where a search index answers faster than a full read. Price and stock come from the live read where your platform exposes them.
When a pull fails
A failed pull changes nothing for the shopper. We keep the last good copy of the catalog and the assistant keeps answering from it. The job retries on its next run, and a repeated failure raises an alert on our side.
The one case a shopper notices is a product added after the last successful pull. The assistant does not know it yet.
What this means for you
Nothing to build and nothing to host. You create the read-only credential once, described on your platform's install page, and we do the rest.
Last updated on