Skip to main content

Product data

Fill missing specs from the web You pick every field

Look products up by EAN, UPC, SKU or brand and model, then merge specs, identifiers, images and prices into your catalog.

Start freeBook a demo
  • 6 identifier types
  • Up to 10 allowed or blocked domains
  • 19 fields, individually switchable

Key features

  • Six ways to identify a product

    EAN, UPC, model number, SKU, brand plus product name, or brand plus MPN. Pick the one your catalog actually carries.

  • You choose the sources

    Leave the web open, or hand research a whitelist or a blacklist of up to ten domains.

  • You choose the fields

    Nineteen fields across content, commerce and compliance, each with its own switch, and existing values stay put unless you say otherwise.

Why it exists

Half a record is not a listing

A supplier file usually carries a name, a barcode and a price. Everything a customer compares on lives somewhere else, and someone has to go and get it.

  • Research is tab work

    Finding the dimensions, the material and the warranty of one product means opening the manufacturer page, a datasheet and two marketplaces.

    Result: the same lookup runs across a selection instead of one product at a time.

  • Copied specs lose their origin

    A value pasted out of a browser tab carries no record of where it came from, so nobody can check it a month later.

    Result: research arrives as a proposal on the product, not as a paste.

  • Some pages never answer a server

    Distributor portals, logins and bot checks return nothing to a server-side request, however the URL is written.

    Result: the browser extension reads those pages from the tab you already have open.

How it works

From a barcode to a full record

Three steps, and the middle one runs without you sitting in front of it.

  1. 01

    Pick the identifier and the products

    ~1 min

    Select rows in the product grid, or apply the current filter to everything that matches, then choose which identifier the research should look up.

    The job is queued for exactly that selection.

  2. 02

    Research runs in the background

    You keep working. Progress shows in the Process Tracker, and results are cached for a number of days you set, so a repeat run does not repeat the lookup.

    One finished research result per product, with the fields it found.

  3. 03

    Merge what you allowed

    Auto-merge fills the fields you switched on, and Overwrite existing stays off, so an empty field is filled and a filled one is left alone.

    The gaps close and your own data survives.

What you control

The guardrails are set before research runs

Every part of the lookup is a project setting, not a decision the model makes on the day.

  • Identifier type

    EAN, UPC, model number, SKU, brand plus product name, or brand plus MPN, with one of them set as the project default.

  • Allowed and blocked sources

    Leave the web open, restrict research to a whitelist, or exclude a blacklist. Up to ten domains either way.

  • Fields it may fill

    Eight content fields, four commerce fields and seven compliance fields, each with a switch. Warranty, certifications and the shipping flags start off.

  • Identifier columns

    It can complete EAN, UPC, MPN and model number, and derive EAN from UPC or the other way round, on empty columns only.

  • Pricing strategy

    Match the average or the lowest price found, undercut either, or price above the market, with rounding, a floor, a ceiling and a minimum margin.

  • Images

    Product images come along with the research when you want them, and stay out of it when you do not.

In the product grid

Run it on one product or thousands

Web Research sits in the same enrichment modal as the AI content types, so the selection you already made is the selection it runs on.

  • One product or a filtered selection

    Ticked rows, or every product matching the current filter, including selections larger than one page.

  • Runs as a background task

    The modal closes and the job continues; the Process Tracker holds the progress and the result.

  • Repeatable from an automation

    The same research step runs inside an automation, so imported products are researched without anyone opening the modal.

The identifier type at the top decides how the product is looked up; the field list underneath decides what may be written back.

Browser extension

Research from the page you have open

Manufacturer portals, distributor logins and marketplaces often never answer a server request. The WISEPIM extension runs inside your browser, so it reads the page exactly as you see it and sends the specs, images and prices back to the matching product in your catalog.

  • Works on pages behind a login, a cookie wall or a bot check
  • Lifts specs, images and prices from the tab you already have open
  • Matches what it finds to the right SKU or EAN in your catalog
  • Available for Chrome, Edge and Firefox

Install it in your browser

Add to ChromeAdd to EdgeAdd to Firefox
How the extension works

FAQs

Web Research questions, answered

What it can look up, where it may look, what it is allowed to write, and what happens when a page refuses.

Still have questions?

Can't find the answer you're looking for? Please get in touch with our team.

Contact Support

Six types: EAN, UPC, model number, SKU, brand plus product name, and brand plus MPN. One of them is the project default and you can pick a different one per run. EAN and UPC give the cleanest results because they are globally unique; brand plus product name is the fallback when your catalog has no barcodes.

Yes. Sources can be left open to the web, limited to a whitelist, or filtered by a blacklist, with up to ten domains in either list. A whitelist is the strict option: research then only reads the domains you named, which is what you want when only the manufacturer's own site counts as a source.

Not by default. Overwrite existing is off, so research fills empty fields in the groups you enabled and leaves filled ones alone. The identifier derivations between EAN and UPC only ever write to an empty column, and warranty, certifications, the shipping flags and the long name start switched off because they change more than a text field does.

It can propose one. Pick a strategy: match the average or the lowest price found, undercut either, or sit above the market. Then set rounding, a minimum margin, a floor and a ceiling, and whether the result goes to the price or the special price. Writing prices back is a separate switch, and nothing changes until you turn it on.

Some pages never answer a server request at all: distributor portals behind a login, cookie walls and bot checks. For those, the browser extension reads the page from the tab you already have open and sends the result to the matching product. A product with no public presence returns nothing, and switching the identifier type is usually the next thing to try.

Next step

Close the gaps with public data

Point Web Research at a selection, check what it found, and merge only the fields you switched on.

Start free
Book a demo

Web Research documentation

Choosing an identifier, restricting sources, the nineteen mergeable fields, and how the pricing strategy is calculated.

(opens in a new tab)