Serpstat Keyword Research Guide 2026: Every Report, Credit and Workflow

What the keyword module covers, and what it really costs

Serpstat's keyword research module gives you volume, difficulty, CPC, intent and AI Overview presence across 230 Google country databases, plus Bing for the US. Around the core reports sit Search Suggestions, Search Questions, Top by Keyword, Keyword Trends, keyword lists and clustering, and for US research the autocomplete ideas come from Amazon and Perplexity as well as Google. This guide follows the reports in the order you would use them, as of October 2026, and relies on Serpstat's documentation rather than hands-on use.

Which plan you need

Every paid plan includes Keyword Research ^[1], and keyword clustering is available from Individual, the cheapest paid plan. Keyword lists, batch analysis and the AI tools start on Team. Freemium gives 10 searches a day and 10 rows per report, which is enough to look around but not to build a real list. For most solo researchers, Individual covers the core of the module. The question is not features but allowance, and that is where people get caught.

The daily search budget is the real limit

Each search spends one of your daily searches: 100 on Individual, 500 on Team, 1,000 on Team x2 and 5,000 on Agency ^[1]. Inside Keyword Research, one keyword result costs one credit. Sorting, filtering and switching reports also count against the allowance. That surprises people. A researcher on Individual who opens the Overview, moves to Keywords Selection, sorts twice and applies two filters has already used six searches on a single topic. Plan your clicks before you make them. Researchers who sort and filter all day long should budget for Team from the start.

Qualify every keyword in the Overview first

Pick the country database before you search, because volumes, CPC and competitors all change by market. Depth varies a lot, from 3.7 billion keywords in the US to a few million in small markets, so check yours first. The help center calls the Overview "your starting point for any keyword analysis in Serpstat" ^[2]. You enter a keyword, choose a database and see volume, competition, CPC, difficulty, keyword previews and the domains ranking for it. Use it to decide whether a keyword deserves deeper work at all.

What each number actually tells you

Volume is the average monthly search frequency over the past year, so a seasonal keyword's average can hide a December peak. Keyword Difficulty scores the climb to the top 10 from 0 to 100, built from backlinks, URL complexity and whether a query is commercial. The help center suggests new or mid-authority sites aim below KD 40. Treat it as a guide, not a promise. The score cannot see your own site's authority or content depth, so a keyword at KD 35 may be easy for an established blog and hard for a new domain with no links.

Competition measures paid-search crowding from 0 to 100%, while CPC is the average Google Ads click price. A keyword can show high competition but low KD when few strong sites rank organically, and those commercial terms are worth chasing with content. The intent column labels keywords informational, navigational, commercial or transactional, but only for Google US and UA. Everywhere else, judge intent from the results page yourself. Since June 2025 you can also filter for keywords where your domain is not cited in the AI Overview. That list is a content priority list.

The thirty-second check

The help center's routine comes down to four questions: is there enough volume, is the KD realistic, does CPC confirm commercial value, and who holds the top five. If Amazon, Wikipedia and big brands fill page one, a low KD score will not save you. The competitors list is based on SERP rankings, not business category, so a forum or a news site may be your real rival for a query. That is useful. It tells you which format Google rewards before you write a word. The whole check should take under 30 seconds per keyword.

Build the list with Keywords Selection and Related Keywords

Keywords Selection is the core list-building report. It expands any seed into every variation that contains those words, and each row carries volume, difficulty, CPC, competition, intent and the number of results. The help center's filter recipe is a good default: volume above 100, KD below 50, intent transactional or commercial, sorted by CPC descending. Export the top 50 to 100 and group them by meaning, and each group becomes a page. By default the report also hides misspelled keywords. Remove that filter for a big content hub, since misspellings show how real people type, but leave it on when you plan pages.

Related Keywords finds what the seed word misses

Related Keywords surfaces semantically related terms that may not contain your seed at all. In practice this is where most new ideas come from, because people rarely phrase a need the way a marketer does. Connection strength counts the top 20 pages that rank for both the analyzed keyword and the related one, so higher means closer. The help center suggests filtering above 5 and adding those terms to an existing page, and claims that can double a page's keyword footprint. Test it on one page.

Capture how people really phrase a need

Search Suggestions pulls autocomplete ideas in real time from Google, and from Amazon and Perplexity in the US. Some have no volume yet. Search Questions is a filtered subset that shows only question-form queries, and you can narrow it by who, what, where, when, why or how. The help center maps those words to intent: how for process, what for definitions, why for education, which and best for commercial comparisons. Pick five to ten questions above 50 searches and answer each under its own heading. A list of twenty real questions can turn a thin article into one that wins People Also Ask boxes, for one or two searches.

Amazon, Perplexity and the export block

Serpstat added over 25 million Amazon US and over 16 million Perplexity suggestions in February 2025. They are strong for buyer phrasing. They are also US-only, and export and bulk export are not available when Amazon or Perplexity is selected ^[3]. Copy what you need by hand, or switch back to Google to export. AI Search, added in June 2025 and available to all users, works from the other direction: it finds keywords that are semantically close in meaning even when phrased differently. It helps most on new topics and long-tail terms where exact-match reports return little. Read its output critically.

Size the competition before you write

Top by Keyword shows the complete top 100 for any keyword, with each page's position, estimated traffic and keyword count. It is a snapshot from Serpstat's latest crawl, not a live feed. The help center's routine for finding a beatable page is simple: look at positions 5 to 10, find pages with fewer than 20 keywords, and check their domain's SDR. A thin page at position six to nine on a low-authority domain is your best target. Beat it with a better page and modest link building.

Keyword Gap and missing keywords

Keyword Gap, renamed in October 2024, compares domains or URLs and has a Missing column for keywords competitors rank for and you do not. It turns competitor research into a to-do list. For single pages, the URL analysis reports show missing keywords against competing URLs, and the Google Sheets add-on runs the same report for a list of URLs at one API request per URL. Run it on your top 50 pages each quarter. On the paid side, the Overview's panel shows keyword variations used in active ad campaigns, and high-CPC terms that competitors pay for are often worth an organic page.

Keyword lists, trends and the credit math

Keyword lists collect keywords from any report or from a file and show their metrics across several search engines at once. They start on Team, with 10,000 keywords per list, and feed Rank Tracker and clustering. A volume update costs keywords times search engines times 2 credits, so updating 1,000 keywords across 2 engines costs 4,000 credits, and each keyword comes back with a 12-month trend graph. Run it before seasonal planning, not every week. Difficulty bands, Google only: up to 20 is easy, 21 to 40 medium, 41 to 60 difficult and above 60 very difficult. A new site should stay in the first two bands for its first six months.

Keyword Trends

Keyword Trends shows trending Google queries by country and city, with data available for up to one month, and sends keywords straight to Rank Tracker, clustering or batch analysis. It draws on the website and backlink analysis pool, not tools credits: each selected base costs 2 credits to view and each exported row one credit, so viewing five countries costs ten credits. Export only the rows you will use. News sites and local businesses gain most. Speed is the whole edge.

Clustering: from a keyword list to a site plan

Clustering groups keywords by the URLs they share in Google's top 30, and it takes 3 to 50,000 keywords per project ^[4]. The help center says grouping 1,000 keywords by hand takes two to three days. Clustering takes minutes. Strength sets how many shared top-30 URLs make two keywords related: Weak means 3, Medium 8 and Strong 12. Soft clusters need one qualifying pair, while Hard clusters need every pair to qualify. Use Strong and Hard for close synonyms and product variants, and Weak and Soft for a broad store or for building ad groups.

What clustering costs

The pricing rule is plain: "Each keyword costs 5 credits: 20 keywords = 100 credits." That means Individual's 5,000 tools credits cluster about 1,000 keywords a month, and Team's 25,000 cluster about 5,000. A 5,000-keyword project eats the whole monthly Team allowance, so large sites should clean and deduplicate their lists first and cluster in stages. Changing settings without adding keywords costs nothing. Experiment freely. Each finished cluster shows connection strength and homogeneity scores from 0 to 100 and a Metatop of the most frequent competing pages, which are your benchmarks. Unsorted keywords either need their own page or a looser run.

Exports, bulk work and automation

Reports export in six formats, and finished exports stay in Latest reports for 60 days. Export the filtered list, not the raw one. Each report also caps rows at 2,000 on Individual and 10,000 on Team, so for a large retailer whose competitors rank for hundreds of thousands of keywords, Individual shows only a sliver of the picture. Serious ecommerce research starts on Team or Team x2. The monthly export allowance scales the same way:

Batch analysis, also from Team, checks long lists against several databases at once, up to 100,000 keywords or domains, and can collect variations from four reports in up to five regions. The API's keyword methods return volume, difficulty, CPC, related keywords, suggestions and top-100 results, and the daily search limit does not apply to them, so bulk research is faster there. The Google Sheets add-on runs the main reports for a list at one API request per keyword. Our Serpstat API guide covers the setup.

Four workflows that put it together

For a new page, qualify the topic in the Overview, run Keywords Selection with the recipe filters, add Related Keywords above strength 5, and pull Search Questions for headings. Cluster the result and write one page per cluster, using the Metatop pages as benchmarks. For an old page, find its ranking keywords in Site Analysis, run Related Keywords on its main term with the same strength filter, and add the missing terms to headings and FAQs. The help center expects five to ten related terms to grow a page's keyword footprint sharply.

For a product category, use Search Suggestions with the Amazon source for buyer phrasing, Keywords Selection with commercial intent for head terms, and Strong-Hard clustering to split variants into pages. Ecommerce teams get more from the Amazon toggle than from any other single feature. For PPC, sort Related Keywords by CPC to find commercial terms you are not bidding on, check PPC competitors and ad examples for wording, then cluster with Weak-Soft settings to build ad groups. Same tools, different order.

How it stacks up against Semrush, Ahrefs and Mangools

Semrush's SEO plan includes basic keyword research, with fuller keyword tools from Starter, and its FAQ gives SEO users 3,000 Analytics searches a day. That is thirty times Serpstat Individual's 100, so a full-time researcher who lives in reports may find Semrush cheaper per hour of work despite its higher price.

Ahrefs includes Keywords Explorer from Lite but lists keyword clusters and search intents among its Standard features at $249, while Serpstat includes clustering from Individual at $69. For clustering on a budget, Serpstat wins clearly. Mangools sells KWFinder bundled with its other tools, from $49 a month on Basic. Serpstat Individual costs $69 and adds clustering, Amazon and Perplexity suggestions and AI Search. Mangools is simpler. Serpstat goes deeper.

Freshness, coverage and the mistakes to avoid

Serpstat's FAQ says a share of keywords is updated every day and that high-volume keywords update faster than low-volume ones, while Search Suggestions are live. For volatile niches, lean on suggestions. Coverage is harder to pin down. The statistics page lists 8.91 billion keywords, while the keyword research page says over 7 billion in one place and 7.53 billion in another, and none of the figures carries a date. We also could not confirm whether intent now covers more databases than the US and Ukraine. Judge coverage by searching your own niche. Then avoid the usual mistakes:

• Searching the US database for a UK site.
• Trusting a low KD on a page full of giant brands.
• Exporting unfiltered lists and burning rows.
• Forgetting that Amazon and Perplexity results do not export.
• Clustering dirty lists full of duplicates.

The bottom line

Serpstat's keyword research is complete, affordable and strongest where rivals are thin: autocomplete from Amazon and Perplexity, semantic AI Search, and clustering from the cheapest paid plan. It suits content teams building topic clusters, ecommerce sites researching buyer phrasing and PPC managers mining CPC. The daily search allowance is the real constraint, so plan around it. Start each topic in the Overview, build the list in Keywords Selection, widen it with Related Keywords and Search Questions, then cluster and write one page per cluster. Send those keywords to a Serpstat Rank Tracker project the day the page goes live.