Keyword research

Type a niche and the page returns keyword and tag ideas for it: primary phrases, long-tail variants, terms the model expects to be rising, and a set of Etsy-ready tags.

What it is, and what it is not

It is a language model writing suggestions from your niche. It is quick, it is good at phrasing, and it will surface wordings you had not thought of.

It is not measured search data. There is no lookup of how many people searched a phrase last month, and the competition reading is the model’s impression rather than a count of competing listings.

That distinction decides how much weight to put on it. Treat the output as a list of candidates to test, in the same way you would treat suggestions from an experienced seller: worth trying, not worth betting a product line on.

If you want signals drawn from real market data rather than generated ones, that is what Market Intelligence is for. The two are complementary, and mixing them up is the mistake worth avoiding.

Judging the output

Some of what a model produces will be excellent and some will be plausible-sounding filler. A few checks separate them quickly:

• Would a buyer type this? Phrases that describe your process (“high resolution digital file”) lose to phrases describing the thing (“boho nursery print”). Buyers search for what they want, not how it was made.
• Is it specific enough to rank? “Wall art” is unwinnable. “Terracotta arch wall art” is reachable.
• Search it on Etsy. Thirty seconds in the Etsy search bar tells you more about real competition than any generated label. Look at how many listings come back and whether the top ones look like yours.
• Does it fit what you sell? A strong keyword for a product you do not make is not an opportunity.

The Etsy search bar is the best free tool here, and it is worth using on any phrase before it goes into a listing.

Using it well

Seed it from a listing. Opening this page from a listing carries the subject across, so you get ideas for that specific piece rather than for a niche you retyped from memory.

Keep the long tail. Long-tail phrases convert better precisely because fewer shops compete for them. They will not bring floods of traffic, and traffic that never converts is not the goal.

Feed the good ones into a listing. Suggested tags can be carried straight into Create, which saves retyping them.

Do not replace a working listing’s tags with these. A listing that already sells has proven its tags against real buyers. See Tag anchors explained for why editing winners is the most expensive experiment available.

Limits

Keyword research counts against your daily AI allowance, shared with the other AI features. If you hit the cap the page says so.

The model has no view of your shop’s history, your sales, or what is happening on Etsy today. Two runs on the same niche can return different phrasings; neither is more correct than the other.

FAQ

Why does it not show search volume? Because we would have to show you a number we cannot stand behind. Measured keyword data is licensed from third parties on terms that do not permit passing it to sellers, so the honest options were suggestions labelled as suggestions, or nothing. Displaying an invented volume was not on the list.

Is the competition level accurate? It is the model’s impression. Use it as a nudge, and check the phrase in Etsy search before acting on it.

The same niche gave me different results. Expected. It generates rather than looks up. If a phrase appears across several runs, that is mild evidence it is an obvious one for the niche, which cuts both ways.

Should I use all thirteen suggested tags? Use all thirteen slots on your listing, but not necessarily all thirteen of these. Mix them with anchors that repeat across your shop.

Does this affect my ranking on its own? No. Tags affect what your listing can be found for. Whether it then sells depends on the image, the price and the reviews.