Fill Parameters — pagination works again with a template filter
When you picked a template on the Fill Parameters page, all products were shown at once instead of 50 per page. Pagination now works correctly with a template selected too.
Every entry comes here straight from the production release — with its date, module and type of change.
When you picked a template on the Fill Parameters page, all products were shown at once instead of 50 per page. Pagination now works correctly with a template selected too.
Availability and stock imports from Crussis, ProgressCycle, Trek, Garmin, Grandus and Levelsportkoncept no longer process each variant on its own; they look up and save variants in bulk, up to 1,000 at a time. For 10,000 variants this takes about 12–30 database operations instead of roughly 20,000, so stock updates run much faster.
Importing products from Grandus now processes them in batches of 100 instead of one at a time, and remembers repeated data such as brands and category paths during the run, so it finishes faster. An error in one product no longer affects the other products in the batch. Product images are stored in a new way that keeps the AI scenes created in Photo Studio. The import's result counts are no longer inflated after an error.
Parameter normalisation crashed when a parameter value mapping had no source value in normalised form. Such mappings are now safely skipped.
The product's Raw Data screen and the feed snapshot window used to load very slowly because all the data from all 10 snapshots was downloaded at once. Now only a lightweight list loads first, and the full content of a snapshot is fetched only when you open it. The initial load has shrunk from over 500 kB to about 2–5 kB. An opened snapshot is remembered for 5 minutes, so reopening it does not download it again.
Two dashboard widgets have a new look. “Product Distribution by Vendor” is now a single horizontal bar split into coloured segments per supplier, with a legend and tooltips on hover; suppliers with less than a 5% share are grouped under “Others”. “Vendor Health” shows a table of all suppliers (name, status, product count, time of the last sync and error) with the upcoming scheduled syncs beside it. On mobile the parts are stacked; on larger screens they sit side by side.
Imports from Origos, Entafix and Levelsportkoncept now load parameter and value mappings once at the start instead of looking them up for every product. This saves around 20 database operations per product — for 500 products the count drops from about 20,500 to about 502 — so the import runs faster.
Imports of Heureka-format feeds and of the Schindler and EMOS suppliers now need 76–93% fewer database operations per product, so they run faster. A product's images and parameters are saved in one go instead of one by one, and the import remembers parameter and category mappings while it processes a batch. The import no longer looks up a category shared by several products separately for each of them.
Bulk AI actions and ERP transfers can now run in the background. If you close the dialog while a job is running, a floating indicator appears in the bottom-left corner showing each job's progress with a button to cancel it. When a job completes, fails or is cancelled you get a notification and the product data refreshes. This works for AI descriptions, translations, parameter extraction, scene generation, image alt text and ERP transfers. A finished job disappears from the indicator after 5 seconds.
Each product now records when it was last successfully transferred to an ERP system. You will find the time among the timestamps in the product detail, with a link to “Transfer Logs”, and in the description of the “ERP Transfer” card. Next to its “Manage” button, that card now also has a new “View Logs” button.
The shop megamenu showed no categories when products existed only in subcategories rather than directly in the parent category. A category's product count now includes products from all of its subcategories. Categories without subcategories appear in the menu as direct links instead of buttons that only respond on hover.
Brand names in supplier feeds (especially from Grandus) could be longer than the system allowed. We raised the limit for a brand's name, code and URL slug from 100 to 255 characters.
The translations card in the product detail now shows coloured status badges for each language (green, orange, grey), so you can see straight away how the translations stand. If no target languages are set up, the card points you to “Configure Languages”. The bulk translation dialog shows translation progress and, at the end, a summary of the result with the number of languages.
If an internal connection drops while feeds are being processed (for example after a server restart or a network failure), processing now resumes on its own. Previously it stayed idle until someone restarted it by hand. Running tasks are wound down properly first, and processing starts again after 5 seconds.
Clicking any category in the shop led to a 404 (page not found) error. The category page now loads the category first and then its products, so it displays correctly.
Suppliers with an empty availability feed address were included in the scheduled availability sync, and the import then failed on the invalid address. The scheduler now skips such suppliers.
The new Transfer Logs page in Settings → Logs lists every ERP transfer. For each transfer you see the configuration name, type, a colour-coded status, the number of succeeded, failed and total products, when it ran and how long it took. You can filter by status and configuration, expand the per-product errors including their identifiers, and page through 20 entries at a time. Running transfers refresh every 5 seconds.
We fixed a critical error in batch feed imports. Products, brands, variants, images, parameters and stock levels are now saved correctly by these imports.
You can now import the B2B feeds of the supplier Levelsportkoncept. The catalogue XML feed brings products, variants, prices and images, while a separate availability XML feed brings stock levels per warehouse. Warehouses are created automatically and the compare-at price is filled in from the B2B pricing.
Batch feed imports now detect when a product code in your company is already used by another supplier's product. The imported product's code is then prefixed with its feed identifier, so the codes no longer clash and the import does not end in an error.
You can now import the B2B XML feed of the supplier Schindler. It brings in products with variants, prices and stock levels per warehouse, and also handles feeds encoded in Windows-1250.
If Photo Studio is not enabled for your company, the product detail no longer shows its controls: the image editing toolbar, the scene button, the lifestyle gallery, or the scene and video dialogs. If you open the bulk Photo Studio page directly, you see an information message with a link to settings instead.
Your shop now fully supports Packeta delivery. In the integration settings you enter the widget API key, the countries and the language. Shipping methods can be linked to the Packeta integration, and when a customer chooses such a method, the checkout shows the Packeta widget for choosing a pickup point.
Customers choose a pickup point at checkout using the Packeta widget. The selected point (name, address and GPS coordinates) is saved with the order, and you will find it in the Shipping card of the order detail.
The new E-commerce → Integrations page connects your shop to external shipping and payment providers. You can add, edit and delete integrations there, with separate settings for each provider. Access credentials are stored encrypted.
We'll walk through your feeds, categories and parameters live.