In large shops (tens of thousands of products) the Shopgrade category text is now written even when the category's products lie beyond the saved product list; and when we can't fetch a category's products from the shop, we now say so honestly („We couldn't fetch the category's products from the shop”) with a „Retry” button and no credit charged, instead of wrongly saying the category is empty.Help:A kategóriánál azt írja: „nem kaptunk terméket” vagy „nem tudtuk lekérni a termékeit”
A Shopgrade product or category text stops with a „generation failed” error less often: if the AI's first answer is malformed, the second attempt now actually runs and saves the text.
The Category survey gives clearer advice: of two neighbouring categories whose names differ only in spelling, the one with more products stays writable, and you get a „Twin category” task to merge them; the „Why” of „Not recommended” rows gives the real reason („Error page” only for actual error pages); and the „only one subcategory” and „no featured products” tasks are now one row each, only for your important pages. A task row you mark as done stays closed: if new categories come up later, they go into a new row with only the new ones.Help:Mit jelentenek a Kategória-felmérés rendrakó teendői?·Miért „Nem ajánlott” egy kategóriám a felmérésben?
More accurate category texts in Shopgrade: vintages (e.g. „2022/24”) are written as a range taken from your real products, item-number and delivery-time fields are no longer used as buying criteria, and when your products have no parameters the text still builds on your own brands and products — on hub pages with an example from every subcategory.