Skroutz feed 2026: what your XML needs so your products get classified correctly.

Skroutz doesn’t ask for specific field names. <name> can also be called <title>, and the price <price> or <price_with_vat>. What it does ask is that every product has the items below, and that no field contains HTML.

“Classified” is Skroutz’s word for a product it has matched to a product in its own catalogue.

In some cases other fields become required too. Fashion items need a colour, sizes where there are size numbers, and the additional images if you show them on your eshop. Weight is needed when your shipping charges are based on weight.

And for the file as a whole: it should declare its real encoding and the time it was created (<created_at>), list each product once, and be compressed as zip or gzip when it’s over 10 MB. If it isn’t built at the moment Skroutz requests it, it has to be rebuilt at least once a day.

Skroutz matches your product with the product in its own catalogue. The two items that help it most are the EAN and the MPN, and the specification asks for them in every category.

The EAN is the product’s barcode. GS1, the organisation that issues barcodes, now calls it a GTIN, and every GTIN identifies one product uniquely. That’s why the same EAN never goes on two products: the black and the white model each have their own.

The MPN is the code the manufacturer gives the product, not your own code and not the supplier’s. For books, the ISBN goes in place of the MPN, and books without an ISBN aren’t classified.

The most common mistakes:

• Excel breaks the barcode. It shows it as 5.20123E+12 or drops the leading zeros. Barcodes have to be stored as text.
• The supplier’s code in place of the MPN. It doesn’t help classification, because nobody else has it.
• A wrong last digit. The EAN’s last digit is calculated from the others. If it doesn’t match, one of the digits was typed wrong.

The title should say everything that sets the product apart: manufacturer, model, colour, version. The more precise it is, the faster the product is classified. The manufacturer goes in the title even if it also has its own field.

The title does not include warranty, shipping, payment method, gifts or offers. Nor does it include the size, which has its own field. An example:

In your feed the title will usually be in Greek; the rules are the same.

Two more rules. If the product isn’t new, the title has to say so (used, refurbished, ex-display), or the product is deactivated. And if you change the title so that it describes another model under the same ID, the product is deactivated too: each ID belongs to one product for good.

The manufacturer is whoever made the product, not the brand it’s made for. A phone case for one brand, made by another company, has as its manufacturer the company that makes it.

The category is written as the full path, from the first level to the last: “Computers > Hardware > Internal Hard Drives”, not just “Hard Drives”. Avoid categories that mix different products, such as “Printers & Scanners”: classification is slower there, because it needs an extra check.

For photos, the specification says:

• Size: the product must be over 1,000 pixels on at least one side, with no margins around it.
• No logo: photos with the shop’s logo or a watermark aren’t allowed.
• Format: AVIF isn’t supported. The link starts with https and should lead to the file itself, not to a script that generates the image.
• Number: up to 15 different photos per product, the additional ones included.
• No “No image”: if a product has no photo, leave the field empty. A “no photo” image isn’t refreshed when you later add the real one.

Photos are downloaded by the Skroutz ImageBot. If your server blocks bots, add it to the list of those allowed.

Availability in the XML has to match what your eshop shows. Skroutz recognises four values, written in Greek as below:

In July 2026 the 12 days of the last value became working days, instead of calendar days.

Quantity is a whole number, and when the product runs out it becomes 0. For products with sizes, the quantity is the sum of all sizes: 10 Medium, 15 Large and 5 XL make 30.

Many cancelled orders aren’t caused by a mistake in the XML, but by an XML built once a day while stock changes every hour. When the feed is built from the current data of the eshop or your business software, Skroutz always sees the right stock.

The size field is only for clothing and shoe sizes and anything else with standard sizes. Dimensions or colours don’t go there. You list the sizes that are in stock, separated by commas: “38,39,40 1/2” or “S, M, L”. Clothes and shoes without a size aren’t classified.

Colours don’t go into one product. Each colour is sent as a separate product, with its own ID, and its colour must be the one shown in the photo. If your eshop runs on WooCommerce, Skroutz suggests IDs in the form “product code-colour code”.

For more precision there are size variations (<variations>): one <variation> per size, with its own ID, availability, size and quantity, and optionally price, MPN, EAN and outlet. Watch one detail: for variations, the specification gives different availability ranges from the product (4 to 10 and 10 to 30 days).

During 2026 Skroutz added eight fields to the specification. All of them are optional, and none stops your products from showing if it’s missing. But they give Skroutz more information about your products.

Specifications are written like this: <spec name="Υλικό">Καμβάς</spec> (material: canvas), inside a <specifications> element. Don’t repeat there anything that has its own field, such as manufacturer, MPN, colour or size: it’s ignored. If the same name appears twice, the last value is kept.

The wholesale price is the price at which you buy the product. It’s optional, and whether you send it is up to you.

• The same price for everyone: the price in the XML can’t come with conditions, such as “with a coupon” or “for members”. If it doesn’t include VAT, Skroutz adds the top rate.
• Shipping only if it’s fixed: the shipping field is filled in only when the cost doesn’t depend on weight, region or other conditions. If the amount doesn’t hold for even one product, shipping is switched off for every product in the XML.
• Access for the SkroutzBot: the XML is downloaded by the SkroutzBot from specific IP addresses, which Skroutz publishes. If you use Cloudflare or another tool that limits requests, make an exception for them. The SkroutzBot doesn’t log in with a password, and it needs a valid https certificate.
• One product, once: if your eshop is in two languages, each product goes into the XML once, preferably in Greek. If an ID appears twice, the first one is kept.

Our free XML feed check reads your file in your browser, without uploading it anywhere, and shows you in a few seconds what’s missing or wrong: required fields, price and VAT format, availability, the EAN’s last digit, duplicate IDs, the manufacturer in the title. It also checks the rules of BestPrice and Shopflix, two more Greek marketplaces. It doesn’t check the size of the photos or whether the links open.

Skroutz also has its own Data Feed Validator, which checks the XML against its own rules.

If your XML comes from an eshop plugin that doesn’t give you these fields, or if you sell on more marketplaces, see how we build feeds from your eshop or Odoo that update by themselves, in marketplaces & XML feeds. If your products come from suppliers, first read the five pitfalls of importing supplier XML.