Card description and filling with a neural networkA marketplace product card: how to create one with AI

Description and card fields with AI

A card's text does two jobs at once: a buyer reads it, and the platform's internal search finds you by it. A title and description missing the right words means you simply don't show up for the search that people use to look for your product. So the text doesn't start with pretty phrasing — it starts with what people actually type into the search box.

Search queries first, text second

Guessing keywords is pointless: "thermal mug", "coffee thermos" and "mug with a lid" represent different demand and a different number of buyers. This is where query research earns its keep: you can see how people actually name your product and which phrasing is more popular.

This is exactly what sets Twelver apart: keyword research is built right into the chat. No need to open a separate tool and copy words over by hand — ask in the conversation what people search for in your category, get the search volume, and immediately ask it to build a title and description from those words. Keyword research and copywriting happen in one window.

What to build the title from

  • Main query. What most buyers actually call the product.
  • Key specs. Volume, material, color, size — what people add to the query.
  • Qualifiers. "For whom / for what": a gift, for the office, a gift for him.

Don't turn the title into a keyword dump: platforms rank down cards that stuff keywords, and the buyer just won't finish reading it. Put the main query first, and the rest in natural language.

Description and specs

In the description, AI turns the benefits into flowing text and weaves in the queries that didn't fit the title. Filling in the specs is its own bit of busywork you can also hand off to the model: dimensions, materials, features from a list. But check every value against the actual product — a wrong spec on a marketplace costs you a return.

“A keyword-driven request for your card text”

A ready-made template: how to gather search queries and turn them into a title, description and specs in one conversation.

Guest
2
Account
3
Subscription

Unlocks after a free sign-up

What's next

The same keywords also work on infographic slides. And the requirements for title length and description fields differ by platform — that's covered in the Wildberries and Ozon chapter.

Frequently asked questions

How do I write a product description with AI?
Don't start with pretty phrasing — start with search queries: ask in chat what people search for in your category, get the search volume, and ask it to build a title and description from those words. In Twelver, keyword research is built right into the conversation, so research and copywriting happen in one window.
Can I write a product description for free?
Yes — the first generations are free after signing up on Twelver, enough to build a title and description around real search queries.
What should a card title be built from?
The main query (what most buyers call the product), key specs (volume, material, color, size) and qualifiers like "for whom / for what". Put the main query first and the rest in natural language: platforms rank down cards that stuff keywords.
Can I trust AI to fill in the specs fields?
Yes — the model can fill in dimensions, materials and features from a list, but check every value against the real product: a wrong spec on a marketplace costs you a return.

You can research keywords and turn them into card text in one conversation in the Twelver chat — research and copywriting in one window. The first generations are free after signup.

Try it yourself

Everything in this guide runs inside Twelver

One chat for text, images, video, music and voice — no separate services or subscriptions.

Open Twelver chat
Rate your experience