Product data that marketplaces accept

The problem is not your data. It is that every channel wants something else.

A supplier file has the fields that supplier happens to keep. bol requires a white background and its own category. Amazon caps titles at 75 characters and wants five bullets. Google wants its own taxonomy. Nobody tells you which of your 2.000 products fails which rule — you find out when an upload is rejected, or worse, when a listing quietly underperforms.

Katalogiq is not a second product database to maintain. It is the workbench in between: data comes in from your supplier or your shop, you complete it, and it goes back out in the shape each channel expects.

How it works

Five steps, in the order you actually do them.

Bring your product data in

Upload a supplier file (CSV or Excel) or connect your shop. Messy column names are fine — you map them once and we remember that mapping per supplier for next month. A dry run shows what would be created, updated, skipped or rejected before anything is written.

Say where you sell

Add a channel for bol, Amazon, Google Shopping or your own shop. Pick a ready-made profile and its requirements are already filled in: required fields, text limits, image rules.

See what is missing

One screen per channel: which share of your catalogue cannot be published, and precisely why — this field empty, that photo too small, that title too long. It also tells you which single field blocks the most products, so you know where to start.

Fill the gaps

By hand, with AI proposing values, or by sending your supplier a link that asks only for the missing fields — no account needed on their side. Everything AI proposes waits in a review queue.

Export

A file in the shape the channel expects: right column names, right units, stable image URLs. Only products that pass validation go out, with a list of what was held back and why.

Channel rules are settings, not software

Marketplaces change their requirements without asking. If following that change means waiting for a software release, you lose weeks every time. So the rules live in a setting you edit yourself.

The same product can be publishable on bol and blocked on Amazon. That is not a bug — it is the difference you need to see before you upload.
You decide what each rule costs: blocking, warning, or informational. bol stopped enforcing its image guidelines in July 2025, so those sit as warnings rather than pretending your catalogue is broken.
Every pre-filled number shows its source and the date it was checked, so you can tell a verified figure from an assumption.

Nothing AI invents reaches a marketplace without you

AI is useful for proposing a colour or a material from a photo. It is not something you let write to your catalogue unsupervised — a wrong-but-plausible value does not look like an error, and you find out weeks later.

A proposal below the confidence threshold is never written to the product. It waits in a review queue and nowhere else.
Every value carries its origin: entered by hand, from a feed, from your supplier, or from AI with the model and date.
An export contains only values a human confirmed. You can prove that afterwards, per field.

Photo checks cost nothing

Most image requirements are measurable without AI, so checking your whole catalogue is free. We measure:

Resolution, aspect ratio and file size against each channel's limits.
Frame fill — what share of the photo is actually product. bol wants around 80%, Amazon 85%.
Whether the background is genuinely uniform white, or only white on average (a gradient is caught).
A shadow under the product, and text or logos burned into the image — the rule marketplaces still enforce.

If something could not be measured it says so, rather than showing a green tick nobody checked.

See it on your own catalogue

Upload one supplier file and you will see within minutes which part of it is not publishable, per channel, with the reasons attached.

Create your workspace

Product data that marketplaces actually accept