Free Etsy keyword tool — real autocomplete demand, no invented volumes

Where the demand signal comes from

Two places, both of them Etsy’s own public surfaces, and neither of them a database of search volumes.

The phrases come from Etsy’s search autocomplete — the completions its search box offers while a shopper is still typing. That list is ordered, and the order is the signal: a phrase near the top is one Etsy puts in front of shoppers before the ones below it. Somebody typed those. They are not a model’s guess about what somebody might type.

The crowding number comes from the public results page — how many listings Etsy says it has for that phrase. It is a market-size signal, not a demand one, and the two are easy to confuse: a phrase with a large count is popular with sellers, which is a different fact from being popular with buyers. The pair is what makes a phrase worth a slot. Plenty of shoppers finishing it, not many listings already chasing it.

Both are best-effort, and the tool says so out loud when they fail. These are unauthenticated public endpoints, not an API anybody granted us — so a timeout, a block or an unfamiliar response shape returns nothing rather than something invented, and the report falls back to built ideas with the fallback named on screen. There is a difference the tool keeps hold of, too: an endpoint that answers with zero completions is telling you your seed is too specific, while an endpoint that fails to answer is telling you nothing at all. Those get different sentences, because they mean different things.

For Amazon Handmade, eBay, Shopify, Depop or Walmart there is no signal here at all. That is the honest limit of this page, and it is why the heading says Etsy rather than naming six marketplaces the way our other tools do. None of the six publishes search volumes; Etsy is simply the one this page asks, one phrase at a time. If you want phrases for the others, the tag generator asks each marketplace’s own search where there is one to ask, builds them by rule where there is not, and says which you are looking at.

Why there is no search volume on this page

Because Etsy does not publish one. It has never released per-keyword search counts, to anyone, and there is no API that returns them. Every exact monthly volume you have seen attached to an Etsy keyword — in any tool, including the well-known ones — is a model’s output presented as a measurement. Some of those models are careful. None of them is a count.

We would rather show you a smaller true thing. So the bands here are defined by something you can check yourself in the search box in another tab: a phrase’s position in the list Etsy returns, cut into thirds.

Two consequences worth stating plainly, because they are the cost of being honest rather than impressive. A band only ranks phrases against each other within one search, so comparing a high from one seed with a high from another tells you nothing. And a short list produces coarser bands than a long one, because thirds of five is not the same instrument as thirds of twelve.

That is the whole method. It is also, as far as we can tell, the only page in this category that writes the method down — the comparison with eRank goes through where the numbers in the bigger tools come from.

One seed, worked all the way through

Here is what Etsy returned for soy candle, captured on 2026-09-24, in the order it came back:

And here is that same response put through the report builder this site actually runs — not a screenshot of one, the function itself, on the render that produced this page:

11 of those 21 carry a band, and they are exactly the ones Etsy returned. The other 10 are marked idea and carry no band at all, which is the most important piece of formatting on the page: they are built from your seed rather than observed, and nothing about them says a shopper ever typed them.

Read it the way you would read a shortlist. “soy candles” is the phrase Etsy puts first; it is also the most obvious and therefore the most contested. Further down are the ones with a qualifier attached — an occasion, a material, a recipient — and those are usually where a tag slot earns its keep. A phrase you recognise as describing your actual product beats a phrase with a better band that describes someone else’s.

When the live sources are down, the report does not pretend. The same seed, with both endpoints failing, returns this instead — every phrase built, nothing banded, and a line on screen saying why:

Free with no account, and the only limit is the one that keeps it fast for everybody: about 20 checks per 10 minutes per connection.

Whose search log you are actually reading

Etsy runs one search log and it is mostly English, and we call it from an American address, because that is where our servers answer from. For an English keyword neither fact matters. For a short word in another language they matter enormously — which is what the listing-language picker above the box is for.

It does not translate anything. It names a market. Your keyword goes to Etsy exactly as you typed it; the setting tells Etsy which country’s shoppers to answer about. It cannot invent completions a prefix does not have — nothing can — but where the same letters have been typed by people searching in several languages, it decides which crowd you hear from. Here is the clearest example we have. vela is Spanish for candle:

Both captured 2026-09-24, the same minute, from production. The first list is merchandise from a fantasy series whose fictional city begins with the same four letters — not one candle in it. The second is candles, with a single fandom result still hanging on at the end, which is the honest shape of the thing: naming a market filters, it does not sterilise.

French does the same trick. bague is French for ring. Asked bare from that same American address it returns “baguette ring”, “baguette earrings”, “baguette diamond” and “baguette necklace” — the gemstone cut, not the thing you wear. Asked about the French market it returns “bague femme”, “bague homme”, “bague fiançailles”, “bague argent” and “bague mariage”: a woman’s ring, a man’s, an engagement ring, silver, a wedding. Captured 2026-09-24, same run.

The setting is the language of your listing, not of this interface, and it is the same choice the tag generator and the listing generator use — pick Spanish here and it is still Spanish when you move on to writing the listing. It covers English, German, French and Spanish, because those are the languages we have both a market for and a body of evidence about.

A second word still helps, and the picker does not replace it. Add the material, the scent, the occasion, the recipient — whatever makes the phrase specific to the thing you actually sell. The market setting fixes the case where four letters mean something else to a larger crowd; it does nothing about a seed that is simply too vague to have interesting completions in any language.

One methodological note, since this is the sort of claim people check: that endpoint reads the caller’s geography as well as the query, so running the same comparison from a European laptop produces a quieter result and the wrong conclusion. Every capture above is from production — the same egress this tool calls out of — which is the only place the comparison means anything.

The half of the report we build, not measure

Every report ends with a run of phrases marked idea. They exist so the tool is still useful when Etsy is quiet or your seed is too specific to have completions, and they are assembled from your seed plus a fixed list of modifiers. Deterministic: the same seed gives the same ideas, every time, forever.

That fixed list is in English, and on a non-English seed it shows. For soy candle the ideas read as you would expect — “personalized soy candle”, “handmade soy candle”, “soy candle ideas”. For vela the same machinery produces “personalized vela” and “handmade vela”, which is an English modifier bolted to a Spanish noun and not a phrase anybody would search.

We are telling you rather than hiding it, for the same reason the bands say “estimate”: a tool that quietly hands a Spanish seller half-English suggestions has spent their trust to look complete. Treat the ideas as a prompt for your own vocabulary — the words you would use for the material, the occasion and the buyer — rather than as phrases to paste. The measured half of the report is the half that came from outside this building.

What a phrase is worth once you have one

A keyword is not an achievement; it is an input to two fields with hard budgets, and those budgets decide how many phrases you can actually carry.

The tags field is where most of them live. Etsy gives you 13 of them at 20 characters each, checked against Etsy’s own help centre on 2026-09-03. That cap is tighter than it sounds: plenty of the phrases in a report above are longer than 20 characters and cannot be used whole. The tags guide covers what to do about that, and the tag generator does the trimming and de-duplication for you.

One phrase belongs in the title, and it belongs at the front. Etsy’s cap is 140 characters but its current guidance asks for a short natural title, so writing to about 80 is the sensible target — and the first 40 characters are the part a phone shows and a search engine reads first. Put your strongest phrase inside them and check it did survive: the title checker reads that opening fragment back to you, against whichever marketplace you are writing for.

Then say it once more where it counts. The first 160 characters of the description are what gets pulled into a search snippet, so the phrase should appear there in a sentence a human would read. After that, stop — repeating it further down buys nothing and reads badly. The keyword research guide works through picking between two phrases when both look plausible.

Free here, and what the rest of it adds

The report is the whole thing. It is free, there is no account, nothing is truncated behind a sign-up and no phrase is held back. What the generator adds is not a better keyword — it is the work that comes after one:

What neither column contains is a promise about what happens next. We cannot see your sales, your views, your favourites or where a listing ranks — nobody outside your own seller dashboard can — so no number on this site will ever be presented as one. A phrase with real demand and few competitors is a better bet than one without. It is still a bet.

If you already have the listing and want to know how the words in it measure up, the listing checker scores a live one from its public URL. If you want the phrase turned into the rest of the listing, the generator is where that happens — it arrives with the seed already filled in.

Related tools and guides

• How to find Etsy keywords — Where phrases come from before any tool ranks them
• Tag generator — The same signal, asked about a market, turned into tags
• How Etsy tags work — What the field does once the phrases are in it
• Title checker — Whether the phrase survived into the first line
• Sellura compared with eRank — Where the search-volume numbers in other tools come from

Sellura is an independent tool and is not affiliated with Etsy, Amazon Handmade, eBay, Shopify, Depop or Walmart. The phrases and counts on this page come from Etsy’s public search surfaces; every captured example carries the date it was taken.