Free Keyword Research Tool: 50 Ideas With Real Volume

One seed term in, up to 50 rows out. Each row is a real Google query that contains the words you typed, and each carries three numbers: average monthly search volume, average cost per click, and a competition index between 0% and 100%. The list is sorted by volume, largest first.

That last detail about containment matters more than it sounds. This is a literal expansion, not a semantic one. Seed it with "running shoes" and every idea returned will have "running" and "shoes" in it somewhere. You will not get "trainers", "sneakers", or "best footwear for marathon training", because none of those contain your words. The tool widens a phrase you already have. It does not find the phrase you haven't thought of yet.

Knowing that changes how you use it. The seed is not a search box — it's the ceiling of the result set. Choosing it well is most of the work.

From Google's own advertiser data, for the country and language you selected. These are the same figures someone bidding on the keyword would see, which is both the strength and the catch.

The strength: nobody is modelling anything. A volume of 74,000 is Google's number, not our estimate of Google's number.

The catch is threefold, and no tool of any price escapes it.

It's an annual average. The figure covers the previous twelve months of monthly searches. A term that tripled last month and a term that halved can report the same volume. If you need direction over time rather than size, Google Trends is free and built for exactly that.

It's bucketed. Google rounds to a scale — 1,000, 1,300, 1,600, 2,400, 5,400, 8,100, 12,100 and so on up. Two keywords showing the same number are frequently not equal in reality; they've simply landed in the same bucket. Never pick between two keywords on a volume difference of one step.

It groups close variants. This is the big one, and it gets its own chapter.

One consequence worth naming early: a keyword reporting no volume at all is not the same as a keyword nobody searches. It can simply be too small or too new for Google's advertiser reporting to have bucketed it yet, which is why zero-volume keywords are worth a second look rather than an automatic delete.

Because Google reports one volume for a whole cluster of near-identical queries, and this tool shows you the cluster.

We ran three seeds through the tool on 8 August 2026, all US / English, and counted how many genuinely different volume numbers came back in each 50-row list:

Fifty rows, two numbers. The "running shoes" list contained "shoes for running", "shoes running", "running with shoes", "runners running shoes" and eight more, every one of them reporting exactly 368,000 searches and a $1.08 cost per click. Those aren't twelve opportunities. They're one keyword, and Google is telling you so by giving them one number.

Word order, stopwords, plurals and small function words get normalised away. Google's own systems treat "shoes for running" and "running shoes" as the same intent, so the reporting collapses them, and so should your content plan. Writing one page per variant is the single most common way a keyword list turns into a set of pages that cannibalise each other.

The practical move takes ten seconds: read down the volume column and mark the repeats. Every group of matching numbers is one target. Pick the phrasing in that group that reads most naturally to a human, use it in your title, and let the rest sit in the body where they'll be matched anyway.

What survives that collapse is the interesting part of the list. In the "seo audit" run, the top row was the seed itself at 5,400 — but further down sat "seo audit tools" at 2,900, carrying its own distinct number. A separate number means a separate intent, which means a page that can rank on its own instead of competing with one you already have.

Google Ads competition. How many advertisers are bidding on that keyword relative to others, expressed from 0% to 100%.

It is not a difficulty score, and we deliberately don't publish one. Here's the distinction, because it's the one that costs people the most time: competition tells you how many people want to buy clicks on a term. Difficulty would tell you how hard it is to earn them. Those come apart constantly. A keyword at 100% competition can have a top ten made of thin affiliate pages you could outrank in a month. A keyword at 5% can be locked up by three publications with a decade of links each, precisely because there's no money in it and nobody has bothered to try.

The reason we don't ship a difficulty number isn't modesty. There's no standard for one. Every vendor computes its own 0–100 score from its own link index, so a 34 in one tool has no defined relationship to a 34 in another, and none of them are Google's. A score with no shared definition looks like precision and isn't.

Use the competition column for what it does measure, which is real: commercial pressure. High competition plus a high cost per click means somebody is making money from that query. In the same August run, "email in marketing" reported a $57.28 average CPC. That's a keyword worth more than one with ten times the traffic and no bidders — the advertisers have already done the qualification work for you.

For the organic question, look at the results themselves. Our keyword competition checker scores the pages currently ranking rather than the keyword in the abstract, and a live rank check shows you what's actually sitting in the top ten today.

The sequence we use, in order:

1. Seed with the plainest version of your topic. Two or three words, in the customer's language rather than yours. If your product page says "workforce orchestration platform" and buyers type "employee scheduling software", seed the second one.
2. Set the market before you read anything. Volume and cost per click are both per-market figures, and the defaults are not your market. Run it once per country you actually sell in.
3. Collapse the duplicates. Mark every group of rows sharing a volume number as one target, as above. A 50-row list usually becomes six to fifteen real candidates.
4. Cut on wording, not on size. Of what's left, keep only the phrasings a buyer would plausibly type. A high-volume row that doesn't match how your customer talks is somebody else's keyword.
5. Check the results page for each survivor. This is the step that decides whether the page you write ranks, and it's the one that gets skipped. Search the term. Look at what format Google is serving, whether an AI Overview is sitting on top of it, and whether the current top ten looks beatable by a page you can realistically produce.
6. Record the country, language and date with the numbers. The tool stores nothing between runs, and volumes get restated. Three fields turn a screenshot into evidence that's still readable in six months.

Steps one through four take about five minutes. Step five takes most of the time and produces most of the value.

Four things, and being clear about them is what keeps a keyword list from becoming a bad content plan.

It doesn't read intent. No keyword tool does. The competition and CPC columns hint at commercial pressure, but they can't tell you whether Google is currently answering that query with product listings, a comparison table, a tutorial or a forum thread. Only the results page tells you that, and format mismatch beats on-page optimisation every time.

It doesn't know what's above the results. Volume is not the same as available clicks. Across the keywords we track, an AI Overview appeared on 27.4% of them in July 2026, up from 18.9% in January. On those queries a share of the demand gets satisfied before anyone scrolls. The number in the volume column doesn't move when that happens — what you can win from it does.

It doesn't show seasonality. An annual average hides a Christmas spike and a summer collapse equally well. Anything with a season needs Google Trends beside it, and there are six other ways to spot a rising term before its average catches up.

It can't find what your seed excludes. Because every result contains your words, the tool cannot surface the adjacent concept you haven't named. That's what the results page, your own Search Console query data, and talking to customers are for. Use this tool to size and compare the ideas you have; use those to find the ones you don't.

So the honest scope is narrow and useful: this tool tells you how big a phrase is, what advertisers pay for it, and which of your fifty rows are secretly the same keyword. Run your seed, collapse the duplicates, and take the survivors to the results page.