Free keyword search tools: what each free plan actually gives you

The best free keyword search tool is a pair, not a single app: Google Keyword Planner for Google's own volume numbers, plus a tool with a difficulty score such as Ubersuggest or Semrush's free tier. Budget for three to ten lookups a day before a paywall stops you.

That pair is enough for one post a week. It is not enough for the way research works when an agent is doing the pulling, which means dozens of small queries, repeated on a schedule, against data you already own. The gap between those two jobs is the whole story of free keyword data in 2026.

The free tiers, side by side

These are the limits each free plan shows as of October 2026, checked against the vendor pages and cross-checked with RightBlogger's October 2026 round-up of free keyword tools, which is the source for the caps below.

• Google Keyword Planner no published search cap, but it sits inside Google Ads and wants billing details on the account. Volumes are rounded and the competition column scores advertisers, not organic difficulty.
• Google Search Console no cap at all, but only for queries your own site already appears for, and Google leaves rare queries out of the report.
• Ahrefs Keyword Generator up to 100 keyword ideas and 50 questions per search with no login. No bulk export and no difficulty history.
• Semrush three uses a day without an account, raised to 10 Keyword Magic Tool reports a day (10 results each) once you sign in.
• Ubersuggest three searches or reports a day across the whole dashboard, with no export on the free plan.
• KWFinder five lookups every 24 hours, each returning volume, a difficulty score and up to 15 related keywords.
• AnswerThePublic one search a day without an account, three with one, and exports capped at 30 keywords.
• Bing Webmaster Tools no published cap, Bing search volumes only, with a country breakdown and the top ten pages per keyword.
• Google Trends unrestricted, and it never gives you a volume. Interest is scored 0 to 100 relative to other searches.

Is Google Keyword Planner still free in 2026?

Yes, and it is still the only free place to see Google's own volume data. The catch is the account. Keyword Planner lives inside Google Ads, so an Ads account with billing details attached has to exist before it will show you a single number. You do not have to launch a campaign or spend anything.

Two details matter more than the price. The volumes are bucketed, so a keyword reported at 1,000 searches covers a wide band rather than a precise count. And there is no organic difficulty column at all: the competition figure describes how many advertisers bid on the term, which says nothing about whether you can rank for it.

If you want those numbers through an API instead of a login, Keyword Planner data is served by the Google Ads API's keyword planning service, and that API is gated by access level. Google's own access-level documentation, updated in October 2026, puts Basic Access at 15,000 operations a day, and Basic Access is still an application you have to win. Free data, gated plumbing.

What each free tool is actually good at

Ranking free keyword tools by search volume is the wrong sort. Rank them by the job you are hiring them for, because almost none of them do two jobs well.

• Volume source: Keyword Planner, because the number comes from Google rather than from a reseller's estimate of Google.
• Difficulty check: KWFinder or Semrush's free tier, both of which return a usable difficulty score on a handful of lookups a day.
• Question mining: AnswerThePublic for the shapes people type, then Bing Webmaster Tools for question keywords with demand attached to them.
• Your own demand: Search Console, the only free source that tells you which queries your site almost ranks for.

One rule saves most of the wasted lookups: never spend a difficulty check on a keyword until a volume source or Search Console has told you the term has demand. The reverse order is how people burn three of their five daily lookups on terms nobody searches.

Where the daily caps bite

A cap of three or five queries a day sounds generous until you count what one content cluster actually costs.

Take a single topic with twelve candidate keywords. You check volume on all twelve, check difficulty on the eight that survive, then check volume again on the long-tail variants the difficulty check surfaced. That is 25 to 30 lookups. At five a day it is a week of work, and the first numbers are already a week stale by the time you finish.

Multiply that by four topics a month and the free tiers stop being free in the only currency that matters, which is your calendar. The paywall is not the problem. The waiting is.

This is why the same handful of tools appears in every free-tool listicle while the same people still buy a paid seat. A free tier is a sample of the product, not a research workflow.

Why a free tool and a free data feed are different things

A person opens a tab, types a keyword, reads a table. An agent calls a tool and expects structured rows back, repeatedly, inside a budget. Every free tier above is built for the first case.

Check the agent-facing path before you commit to a stack. The pattern is consistent:

• Ahrefs ships an SEO MCP server for ChatGPT, Claude, Cursor and Copilot Studio, and its own SEO MCP page lists it on the Lite, Standard, Advanced and Enterprise plans only, with 100 to 500 rows per request depending on tier.
• Semrush and Ahrefs both meter their free web lookups by hand, not by request, which gives an agent nothing to budget against.
• Search Console's API is free and uncapped for your own property, which makes it the one genuinely free feed most sites already own.
• The Google Ads API does not bill per call, but it sits behind an approval process and a daily operation ceiling.

Read the pattern plainly: the free web interface is for humans, the free API is your own Search Console data, and everything else is the paid product wearing a free hat for ten minutes at a time.

How do you run keyword research with an agent on a zero budget?

Use the data you already own as the spine, then spend free lookups only on the few decisions that genuinely need an outside number.

1. Pull 90 days of Search Console rows for your domain and keep every query sitting between position 8 and 25 with at least 20 impressions. That filtered set is your real candidate list.
2. Group the queries by the page that already ranks for them. One page per cluster, no exceptions.
3. Take the top five clusters and spend one free difficulty lookup on each. Five lookups fits inside a single day on almost every free tier in the table above.
4. Write for the cluster rather than the keyword, covering the variants your own query data shows.
5. Re-run the same pull 30 days later and diff the positions page by page.

The fifth step is the one nobody does by hand. A monthly diff of your own positions is a repeatable, structured job, which is exactly what an agent is good at and exactly what a free keyword tool cannot do for you.

What can free keyword data not tell you?

Free data is directionally honest and precisely useless in three places.

• Volume on low-traffic terms. Anything under a few hundred searches a month gets rounded, bucketed or dropped. If your strategy is long-tail, you are working from estimates dressed up as numbers.
• Whether a click exists at all. Rand Fishkin's 2024 zero-click study with Datos found 58.5 percent of US Google searches ended without a click, and only 360 of every 1,000 reached a page Google does not own or get paid by. A keyword can have volume and still be a question Google answers on the results page.
• What you nearly rank for. Third-party tools estimate your positions. Search Console knows them, but only for the queries it decides to report.

The practical answer is to treat free volume data as a filter rather than a forecast. You are asking whether a term deserves one of your five daily lookups, not how many visits it will deliver.

The 20-minute weekly loop

This is the version that fits between other work and still compounds month over month.

1. Sunday: export Search Console queries for the last 28 days.
2. Filter to positions 8 to 25 with 20 or more impressions.
3. Pick one cluster to write and one existing page to refresh.
4. Spend up to five free difficulty lookups on the write.
5. Add two internal links from your strongest page to the one you are refreshing.
6. Log the keyword, the page and the starting position so next week has a baseline.

Twenty minutes, mostly reading your own data, with the outside lookups capped by design instead of by a paywall.

When free stops being the cheaper option

Free turns expensive when you need three things at once: bulk rows, an API an agent can call, and difficulty data on more than five terms a day. That is not a marketing failure. It is the point where the tool is doing work you would otherwise pay a person to do.

Near that line, the honest comparison is cost per useful query rather than sticker price. We put real 2026 numbers behind that in what SEO actually costs, and the platforms that hold up best for a one-person operation are ranked in best SEO platforms for solo founders.

If you are weighing an Ahrefs seat specifically, the free-tier maths changes what you are really buying, which the Ahrefs alternatives breakdown covers plan by plan.

And if rank tracking is the piece you are missing, free options are thinner than they look, as this comparison of rank tracking software shows.

Start with what you own. A Search Console export plus five lookups a week beats a paid dashboard you open twice a month, and it gives an agent something real to work with on day one.