Elbuz pulls products straight from a marketplace, online store or accounting system by access keys and updates your base catalog on a schedule, with no file exports or uploads.
When you need this:
- Your products are on Rozetka, Shopify or another platform — one run brings them in with categories, photos and attributes.
- Prices and stock change daily — a scheduled template updates them.
- You sell on several platforms — all products end up in one catalog.
- You edit some cards by hand — the “Do not update” toggle protects them from the run.
Where to start: find your platform in the list and get its access keys.
What catalog import via API can do
- Catalog transfer — names, prices, stock, categories, photos and attributes come directly, as they are at the moment of the request.
- Categories and manufacturers — matched to yours by name; missing ones are created.
- Protecting your edits — the “Do not update” and “Fixed Price” toggles in the product card.
- Orders through the same connection — on most platforms, enabled with a separate toggle.
- Scheduled updates — the catalog updates itself, and every request leaves a trace in the log.
is the right choice when the other side is a system with its own programming interface and you have access to it (keys, a token or an integration login). The largest platforms have separate articles covering their keys and specifics:
| Source | What it imports |
|---|---|
| Rozetka.UA | catalog + orders |
| Shopify | catalog + orders |
| WooCommerce (WordPress) | catalog + orders |
| Amazon | catalog + orders |
| eBay | catalog + orders |
| Allegro.pl | catalog + orders |
| eMAG (RO/BG/HU) | catalog + orders |
| Kaufland (DE/EU) | catalog + orders |
| Wix.com | catalog + orders |
| Etsy | catalog + orders |
| TikTok Shop | catalog + orders |
| OLX | catalog only |
| Horoshop | catalog + orders |
| SalesDrive (CRM) | catalog + orders |
The list in the import window is longer and keeps growing. Integrations that bring in orders only (POS systems, some marketplaces) are covered in the overview of order import via API.
1
Available connection types- Available connection types — every API source is collected in this one list.
What you need from the platform: access keys
Keys are issued by the source system itself: almost every marketplace or platform has a developer or integrations section where you create a key, a token or an app login and password. Some services need a single token, others need several values — the exact list is in each platform's article from the list.
Keep keys as safe as a password and don't share them. If a key is compromised, revoke it in the source account, issue a new one and update it in the template.
How to connect
- Create a template and choose the integration you need from the list.
- Enter the access keys and click “Test connection” (available for most connections; the Rozetka template doesn't have it). The button contacts the platform with the same keys as the run and replies with the store name or the rejection text — the quickest way to find a stray space in a token.
- In the “Select data to import” section, turn on the data sets (which ones exist). Products and orders are separate toggles: turn on only orders and no products arrive, and vice versa.
- Click “Import” when you can check the result right away. On a few products, check names, prices, images and categories, and what happened to the cards you already had (how to find them).
- Once everything works, put the template on a schedule.
One template is one connection to one account. Each platform or account gets its own template with its own keys and schedule.
What gets transferred to the catalog
Via API, the catalog receives the same things as from a file: names, prices, stock, categories, images and attributes — exactly what the source system exposes in its interface. The difference is that the data is always fresh: not a snapshot taken at export time, but the state at the moment of the request.
In the template you turn on data sets: “Products”, “Categories”, “Photo”, “Attributes”, “Options”, “Manufacturers”, “Orders”. You can't pick individual fields: with “Products” on, the name, description, price and stock arrive as one package.
Products aren't duplicated: Elbuz looks for an existing card by the pair “manufacturer SKU + name” (marketplaces usually don't return the manufacturer SKU, so in practice it matches by name). The run overwrites the card it finds instead of adding a second one.
Categories and manufacturers created by the run
A category comes from the source as a name or a full path. Elbuz first looks for your category with the same name and puts the product there; only if it finds none does it create a new branch. That's why duplicates like «Ноутбуки (2)» (“Laptops (2)”) don't appear as long as the names match.
- Action or result
- Decision: the path depends on the answer
- Products without a category go into the service category “Products - source code”, which is created disabled and excluded from exports. Move cards out freely: a repeat run doesn't change their place in the tree. A deleted category reappears empty once the source has uncategorized products again. How to sort them — in the data import overview.
- Manufacturers are added the same way: a new brand from the source goes into the directory marked with the connection. Don't rename such brands: the next run won't find the brand under the new name, will add it again under the old one and reassign the products. Change the brand name on the platform side.
Which document an order becomes
If the “Orders” toggle is on, each order from the platform is created as a “Customer order” document — not posted. This document type doesn't write off stock, and a reservation appears only after posting: importing orders won't change availability in the catalog.
The document status comes from mapping the platform's statuses to yours — with the “Order statuses”, “Delivery statuses” and “Payment statuses” buttons in the “Related settings” section of the template card. An unmapped status gets the first status in the directory — “like new”.
Importing again doesn't create a duplicate: the document is looked up by the order number in the source together with the source code. How an order goes through customer and product lookup — in the order import overview.
What a repeat run does to your edits
A repeat update overwrites everything the platform returned in that run: name, description, price, old price, quantity, availability, brand, SKU, tags and the enabled/disabled flag. Fixed a name or added to a description in Elbuz — the next run brings back the platform's text without warning.
What a repeat run does not touch in existing products:
- Photos. Images are added to new cards only: a photo you replaced won't be overwritten, but new photos from the platform won't reach an existing card either.
- Attributes. Existing values stay. The exception is Rozetka: it adds the ones the card doesn't have yet, without touching existing ones.
- Position in the category tree. Moved a card to your own category — it stays there. The exception is Horoshop: for a card without “Do not update”, it sets the category the product has on the site.
- Page URL (SEO URL). It's set only when the card is created.
To lock a card completely, open it, go to the “Import protection” section and turn on “Do not update”. The run then changes neither the price, stock, availability nor brand of that card.
In direct platform connections, the flag also keeps the name and description; in SalesDrive imports, the feed overwrites the name and description regardless. The same flag is available as the “Do not update (flag)” column in the “Products” window — handy both for filtering and for setting the flag right in the grid.
1
Do not update2
Fixed Price- Do not update — the import leaves this product untouched.
- Fixed Price — the import does not change the product price.
The neighboring “Fixed Price” toggle protects prices only. In direct connections — Rozetka, Horoshop, Shopify, eBay, Wix, Amazon, Allegro, eMAG, WooCommerce, Kaufland, OLX, Etsy, TikTok Shop — the run doesn't overwrite that product's prices, but updates its name, description, stock and availability as usual.
- Action or result
- Decision: the path depends on the answer
Two platforms write to the same card
If two connections pull the same product, they meet in one card: it's looked up by name (and by manufacturer SKU when the source returns it), not by platform. A field keeps the value from whichever run finished last; templates have no source priority.
The “Source: name” label in the card doesn't switch either — it keeps the template that created the card, so you can't use it to see who last wrote the price. The only way to separate two sources is the “Do not update” toggle on the card. General rules — in the data import overview.
What happens to a product that disappears from the source
Direct platform connections work on a “whatever arrived gets updated” basis. A product missing from a run stays in the catalog with its last price and stock: it isn't deactivated, its stock isn't reset, and the log says nothing about it. Remove discontinued items by hand, or they'll go into exports with outdated figures.
A status change on the platform is another matter: the enabled flag comes from the source, and in Shopify a product with a non-Active status arrives disabled. Rozetka fetches card statuses from its seller account in a separate pass (with errors, invalid, under moderation, hidden by a moderator, outside the allowed price range) and logs the count per group.
The “Delete entries in the base catalog that are not in the imported file” checkbox in the “What to import” section only really works where the catalog comes as a feed — in SalesDrive.
Products are physically deleted with their descriptions, attributes, photos, questions and reviews; so are categories missing from the feed. The SalesDrive form has no source filter, so items added by hand or by another connection go too. Turn it on only when this source is the only thing filling your catalog. More details — in the data import overview.
Scheduled updates
You put the template on a schedule, and Elbuz contacts the source by itself: new products are added, prices and stock are updated. Frequency ranges from several times a day to once a day; large marketplaces limit request rates, so too frequent a schedule runs into their limits.
The schedule works as long as the connection to the source is alive. If the marketplace is unavailable or the keys have expired or been revoked, the next update won't happen, and you'll see it in the log. So keep an eye on the log for the first few weeks.
Two scheduled imports don't run at the same time: the second one logs “Scheduled import canceled due to another template being imported” and waits for the next trigger. The lock of a run that died without writing to the log is released automatically (how).
- Action or result
- Decision: the path depends on the answer
At the end of the run, Elbuz recalculates fields using your formulas, applies markup rules to new products and updates their search data. The catalog doesn't go to your storefront or platforms by itself: that's the job of a separate “Data export” with its own template and schedule (when imported data moves on).
Disable a template or delete it
The “Active” / “Disabled” badge in the card header takes the template off the schedule; keys, settings and imported products stay. The “Import” button works on a disabled template too: to make sure the connection changes nothing, remove the keys or protect the product cards.
1
State pill (Active)2
Import button3
Run history icon- The state pill shows whether the template is Active or Disabled; clicking it switches the template.
- The Import button starts the import for this template.
- The Run history icon opens the list of previous runs.
A deleted template can't be restored, and you'll have to enter the keys again; its schedule tasks are deleted with it. Products, categories, manufacturers and the source label in cards stay. System templates can't be deleted. More details — in the data import overview.
How to check a run and fix an error
The run log opens in the Operations Center — the right-hand panel that doesn't block your work, so you can close the template card.
- The catalog arrives in pages, with a line for each: “Shopify: catalog page 3, products received: 75”.
- At the end comes a summary: “catalog imported, products: 240”.
- Then “Updating values based on formulas”, for some connections “Data integrity check” and “Product indexing”, and the last line “Data import completed!” with “Time spent”.
- A platform error is shown in red, with the reason text from the platform; the operation card in the Center closes as failed if there's at least one such line.
In the “Products” grid, the update is confirmed by the “Update date” column (set on every changed card), “Source: name” (the template that created the card) and “Source: code” (the platform code).
The “Import date” and “Time spent” columns in the template list update even when the platform returned an error, so judge success by the log. For a longer history, use the “Run history” button in the template card header: “Runs in period”, “Average duration”, “Longest run”, “With errors”.
If the connection broke off mid-run
- Products are written to the catalog in batches of 500, as soon as a batch fills up. A break splits the result in two: what arrived is updated, the rest keeps old data.
- An immediate retry is safe: the run overwrites the same items and picks up the rest, with no duplicates in the catalog or in orders.
- If the request limit is exceeded or the server returns an error, the request is retried three or four times, depending on the platform, with a growing pause; only then does the run stop with an error. Elbuz waits 60 to 300 seconds for a single request.
- There's no cap on the number of items. Data is fetched in pages of 25–1,000 items, and some connections pause between pages to stay within limits.
If you connected the wrong thing
To find products the connection created, filter the “Products” window by the “Source: name” or “Source: code” column; then select the rows and delete them in bulk — Elbuz asks “Delete the selected products?”. The filter works even after the template is deleted.
Cards you already had that the connection only overwrote won't show up in this filter — they keep their old source. Filter them by the “Update date” column instead.
There's no way to restore old values in bulk — Elbuz has no product edit log. Only the price is stored separately: if price history is turned on, you'll see the previous one in the product card, in the “Product price history” block. Bulk actions for fixing things — in the data import overview.
Scenarios
Moving a catalog from a platform to Elbuz for the first time
Task: your products already live on Shopify, Rozetka or another platform, and you want to manage the catalog in Elbuz.
What to turn on: a template of the right type, keys in the form and the “Test connection” button; in the “Select data to import” section — “Products” and, if you need images, “Photo”; then “Import”.
Result: each product becomes a base catalog card marked with its source. If a product's name matches an existing card, the run overwrites that card instead of creating a new one. Rename a product on the platform, and the next run creates a new card; the old one keeps its last figures. Nothing reaches the storefront until “Data export” runs.
I set the price myself, stock comes from the platform
Task: you need quantity and availability from the platform, but you calculate the price in Elbuz.
What to turn on: in the product card, the “Import protection” section, the “Fixed Price” toggle.
Result: a direct connection run updates stock, availability, name and description, and leaves your prices alone. “Do not update” doesn't fit this task: it would freeze stock too.
I maintain a few cards by hand
Task: you've rewritten some products your way — name, description, attributes — and don't want the run to bring back the platform's text.
What to turn on: the “Do not update” toggle in the card's “Import protection” section; for many cards at once — the “Do not update (flag)” column in the “Products” grid.
Result: a direct connection skips such a card entirely. Stock from the platform won't reach it anymore — you'll have to maintain it yourself.
I sell the same product on two platforms
Task: the same assortment is listed on two platforms or in two accounts, and both need to be combined into one catalog.
What to turn on: a separate template for each connection, with its own keys and its own schedule.
Result: products with the same name meet in one card with the values of whichever run finished last (why). “Do not update” closes the card to both connections at once — you can't choose which one writes.
What to do when the catalog doesn't import
| What you see | Why | What to do |
|---|---|---|
| A red platform error line or a rejection on “Test connection” | The key was entered with a stray space, has expired or been revoked, or doesn't have permission to read products | Check the key and its permissions in the source account (keys) |
| Orders arrive, but there are no products | Only the “Orders” data set is on | Turn on “Products” (how to connect) |
| “Scheduled import canceled due to another template being imported” | Another template was running at that moment | Nothing: the import runs at the next scheduled time (schedule) |
| Errors about exceeding the request limit | The schedule is too frequent for the platform's limits | Run the template less often |
| Your name or description reverted to the platform's text | A repeat run overwrites everything the platform returned | Set “Do not update” on the card (protecting edits) |
| A discontinued product still shows old figures | Direct connections don't touch products that weren't in the run | Remove or disable the card by hand (disappeared product) |
| The catalog has new data, but the site shows old data | Import doesn't deliver data to the storefront | Check the “Data export” template and its log (where it goes next) |
FAQ
Why is API better than a file?
You don't have to export and upload anything by hand: connect with keys once, put it on a schedule — and the catalog updates itself, directly from the source.
Can one connection pull both products and orders?
Yes, if the integration supports it: the template has separate “Products” and “Orders” toggles (which platform does what).
What happens to products if I delete the template?
They stay in the catalog along with the source label. Filter the ones you don't need in the “Products” window by the “Source: name” column and delete them in bulk.
My platform isn't on the list?
Check the list right in the import window — it's longer and keeps growing. If there's no API integration, import the catalog from a file: in a ready-made format or with universal import.
How many platforms can I connect at once?
As many as you like: each account gets its own template with its own keys and schedule, and they all fill one base catalog.
Where to find it and who has access
- Menu
- Directories → Operations → Data import
- Card protection
- product card → “Import protection” section
- Access right
- the “Import” window in “Group permissions”
- Prepare
- access keys from the platform account with permission to read products
Related topics
- If you also need to pull orders — Order import via API.
- If you want to understand the general mechanics of import, breaks and rollback — Data import overview.
- If the source has no API but has a file — Universal import.
