Professional tool for SEO specialists

Keyword
Clustering Tool

Automatically group thousands of keywords into semantic clusters. Optimize your site structure and content strategy in seconds.

Keywords
50K+Keywords
Clustering Methods
8+3Clustering Methods
Free
100%Free

Keyword Clusterizer

8 grouping methods: shared words, rare-word weighting, semantics, structure, topic density and spelling variants

0 / 500

💡 Any format supported: with or without frequencies, any separator (tab, semicolon, space, etc.)

Clustering Algorithm
Premium Methods
AI Clustering
Unlock unlimited access for AI clustering

Compare all and select best

When unsure which to choose

Less = more keywords per cluster

More = fewer final clusters

Remaining clusterizations:1 / 1

Keyword clustering: a practical guide

What is keyword clustering?

Keyword clustering is the process of grouping search queries that can be answered by the same page. Instead of writing one article per keyword, you plan one page per cluster, so every page targets a coherent topic and pages on your site do not compete with each other.

A good cluster shares both meaning and search intent: “keyword grouping tool” and “keyword clustering software” usually belong together, while “what is keyword clustering” often needs a separate explanatory page.

Semantic, SERP-based and AI clustering

SEO Clusterizer offers several ways to group keywords. They answer the same question — which queries belong on one page — using different evidence.

Clustering approaches available in SEO Clusterizer
ApproachEvidence usedBest for
Lexical and semantic methodsShared words, word forms and term weighting inside your keyword listFast first pass on large lists, no external data needed
SERP-based clusteringHow many ranking URLs two keywords share in Google resultsFinal page mapping when intent matters most
AI semantic methodsMeaning of the query as understood by a language modelSynonyms and phrasing that share few words (paid plans, smaller batches)
  • Lexical/semantic grouping runs in seconds and works for any language list.
  • SERP clustering reflects what Google actually ranks today, so it is slower and uses search data.
  • You can run several methods on the same list and compare the results side by side.

Example: from raw keywords to a content plan

A shortened example of how a raw list becomes clusters and then pages:

Illustrative keyword clustering output (example data, not real volumes)
Cluster (page)Keywords in the clusterSuggested page type
Keyword clustering toolkeyword clustering tool, keyword grouping tool, cluster keywords onlineTool / product page
What is keyword clusteringwhat is keyword clustering, keyword clustering meaning, keyword clusters seoExplainer article
How to cluster keywordshow to do keyword clustering, cluster keywords in excel, manual keyword groupingStep-by-step guide

Manual grouping vs automated clustering

Manual vs automated keyword clustering
Manual (spreadsheet)Automated (SEO Clusterizer)
Time for 1,000 keywordsHours of sorting and filteringSeconds to a few minutes
ConsistencyDepends on the person and the daySame settings give the same result
Intent checksRequires opening SERPs one by oneOptional SERP overlap comparison
Final decisionHumanHuman — the tool proposes, you review

How to cluster keywords in four steps

  1. Collect keywords from your research tool, Search Console or a CSV export; search volumes are optional and are summed per cluster.
  2. Paste or upload the list and pick a method — the tool can recommend one based on the list size and language.
  3. Review the clusters: merge near-duplicates, split groups with mixed intent, rename the main keyword.
  4. Export to CSV, JSON or TSV and map every cluster to an existing or new URL.

What to do after clustering

Validate important clusters with SERP clustering, write or audit the page with the SEO content analyzer, and compare your draft with ranking pages in The Force. For large sites, check many URLs at once with the Bulk Analyzer.

Limits per run depend on your plan — see pricing for current numbers. Ready-made workflows are collected in solutions.

What is keyword clustering?

Fundamental SEO technique for creating effective site structure

Definition

Keyword clustering is the process of automatically grouping semantically similar search queries into logical groups (clusters). Each cluster represents one topic corresponding to one page on your site. This allows you to create content that maximally matches user search intent and Google algorithm requirements.

1000+
Keywords
Clustering
50
Site Pages

Why is it important for SEO?

  • Avoid cannibalization — one page per intent
  • Increase content relevance for Google
  • Optimize internal linking structure
  • Effective content strategy planning
  • Increase organic traffic by 30-50%

How does our tool work?

We combine several independent grouping approaches: rare-word weighting, shared-word similarity, word-fragment comparison, spelling-variant tolerance, semantic context analysis and hierarchical topic building. The system automatically compares the results of all methods and recommends the best one for your data.

Clustering Algorithms

8 advanced NLP techniques for the most accurate grouping

01

Rare-word weighting

Terms are weighted by how rare they are: unique modifiers influence grouping more than generic words.

02

Word-fragment matching

Queries are compared not only by whole words but by short fragments, so different forms of a word stay together.

03

Spelling-variant tolerance

Typos, alternative brand spellings and transliteration do not break a cluster apart.

04

Hierarchical Clustering

Agglomerative algorithm with average linkage for optimal cluster merging.

05

Semantic Analysis

Word co-occurrence matrix for detecting semantic relationships between terms.

06

Meaning proximity

Every cluster keeps its own meaning centre, and new queries join the closest one by meaning.

07

Multilingual Stemming

50+ stemming rules for Ukrainian and English with morphology support.

08

Stop Words

Automatic removal of 100+ function words for both languages for cleaner analysis.

How it Works

Four simple steps to perfect site structure

  1. 01

    Paste Keywords

    Copy your keyword list from Excel, Google Sheets or any other source. Each keyword should be on a new line.

  2. 02

    Choose Method

    Basic method is fast and effective. Advanced creates more precise subgroups using spelling-variant matching.

  3. 03

    Get Clusters

    The system will automatically group keywords by semantic similarity and show relevance level for each word.

  4. 04

    Export Results

    Download ready clusters in CSV or JSON format for further work in Excel, Google Sheets or your CMS.

Clustering algorithms

Choose method based on data size and needed precision

01

Basic Method

Fast grouping with combined similarity metric

  • Shared-word similarity
  • Word-fragment comparison
  • Spelling-variant tolerance
  • Shared word ratio
Speed:Very fast
Accuracy:Basic
02

Rare-word method

Cluster meaning centres with rare-word weighting

  • Rarity-based weighting
  • Meaning proximity
  • Incremental centre updates
  • Combined metric
Speed:Fast
Accuracy:High
03

Semantic Method

Word co-occurrence analysis in context

  • Co-occurrence matrix
  • Context window ±2
  • Semantic similarity
  • Soft group merging
Speed:Medium
Accuracy:Very high
04

Hierarchical Method

Bottom-up agglomerative clustering

  • Soft group merging
  • Dynamic merging
  • Optimal groups
  • Flexible threshold
Speed:Slow
Accuracy:Maximum
The full SEO stack

From keyword clustering to a complete SEO stack

Free clustering is only the entry point. Four premium areas cover the rest of the semantic workflow: AI-driven grouping, page quality analysis, competitor text analysis of the Google Top 10, and clustering based on the real search results.

AI semantic clustering

A language model reads search intent and niche context where pure math falls short.

Content Analyzer

20 page quality factors, a score, and an automatic content brief in Markdown.

The Force — Top 10 analysis

Full text analysis of every page ranking in the Top 10 and the terms you are missing.

SERP clustering and data

Grouping by real SERP overlap plus Volume, CPC and KD from Ahrefs.

AI semanticsPro and Agency

AI semantic keyword clustering

Algorithmic methods group queries by shared words and statistics. That is fast, transparent and works well on large, uniform keyword sets — but it does not understand meaning. "Buy running shoes" and "marathon footwear" look different mathematically, yet Google treats them as one page.

The AI mode sends your keyword set to a language model that evaluates search intent, query type (informational, commercial, navigational, transactional) and topical proximity. The resulting cluster reflects what a person is actually looking for, not how many characters match.

Two specialised profiles are available on top of that: Google SEO for classic informational and commercial projects, and iGaming for a market where brands, providers, bonus mechanics and geo modifiers create semantics that generic models handle poorly.

  • Intent instead of word overlapSynonyms, transliteration, jargon and different phrasings of the same need land in one cluster.
  • Google SEO profileClusters follow organic search logic: one page per intent, without artificial fragmentation.
  • iGaming profileCorrectly separates brands, providers, game types, bonus mechanics and geo queries.
  • Smart RecommendationThe system profiles your keyword set — size, query length, repetition — and suggests the method that will perform best.
  • Compare methods side by sideRun several algorithms on the same set and compare the cluster distribution before you commit.
Try clusteringSee plan limits
AI runs per day on Pro / Agency
20 / 100AI runs per day on Pro / Agency
keywords per AI run
500keywords per AI run
specialised niche profiles
2specialised niche profiles
AI modes

How AI clustering works — in detail

AI modes are not a magic button. They are a controlled pipeline where the language model handles exactly what algorithms cannot: meaning, intent and niche context.

  1. 1. Preparation and normalisation

    Your keyword set is de-duplicated, cleaned and normalised while search volumes are preserved. The model receives compact batches instead of a raw file — lower cost, more stable output.

  2. 2. Meaning, not words

    The model judges what a query is actually about: "buy washing machine" and "order washer cheap" belong to one page even with almost no shared words. This solves the core weakness of classic methods — synonyms, slang and different phrasings.

  3. 3. Intent detection

    Every query gets an intent type: informational, commercial, transactional, navigational, comparative. Clusters never mix "what is X" with "buy X", so your site structure is correct from the first run.

  4. 4. Niche modes

    AI Google SEO follows search-result logic and page types. AI iGaming understands brands, providers, bonuses, geo and local niche specifics that generic algorithms ignore.

  5. 5. Naming and priorities

    For each cluster the model produces a human page title, a primary keyword and secondary ones. You get a ready heading for the brief, not "Cluster #17".

  6. 6. Validation and self-correction

    A control pass splits over-broad groups, merges near-identical small ones and isolates outliers. Every keyword ends up in exactly one cluster.

Understands synonyms and slang

Queries with zero shared words land together when they target the same page.

Separates by intent

Informational and commercial queries stay apart — less page cannibalisation.

Ready page titles

Each cluster arrives with a title, a primary keyword and brief-ready logic.

Controlled quality

A second model pass removes groups that are too broad or too fragmented.

Transparent output

You can see why a keyword landed where it did, and export the full distribution.

Your data does not train models

Keywords are processed within the session and are not used to train third-party models.

AI modes are available on paid plans due to compute cost. For very large sets, run an algorithmic method first and apply AI to the priority segments.

Process

How we cluster — from a raw list to a site structure

A high-level view of the process: what happens to your keyword set at every stage and what you get at the end.

1. Normalising the set

The list is cleaned of duplicates, stray spaces and stop words; the language and the volume column are detected. Different forms of the same word are reduced to a common base so that "buy laptop" and "laptops buy" never end up in separate groups.

2. Measuring closeness

Every pair of queries gets a closeness score from several independent signals at once — shared words, rare modifiers, spelling and usage context. No single signal decides alone: the final score is their balance.

3. Forming clusters

Queries gather around meaning centres. A cluster that nothing else fits into is closed; one that grows too broad is split into subtopics. Isolated queries are not forced into a group — they are kept aside.

4. Refinement and validation

The system cross-checks the groups: merged topics are separated, split ones are reunited, and each cluster gets a head query and total demand. Several methods can be run in parallel so you can compare outcomes.

5. A ready structure

The output is a cluster table with the head query, demand volume, keyword count and each query's relevance to its group — an immediately exportable blueprint for your site pages.

Example output

A fragment of a real 42-query set after clustering

Main commercial page

online casino uk

12 queries · 62,800 searches/mo

Generic commercial phrasings with the same intent — one landing page.

"Best of" roundup

best online casino uk

9 queries · 14,200 searches/mo

Comparison intent — a separate ranking or roundup page.

New casinos

new online casino uk

6 queries · 6,570 searches/mo

Freshness intent — a page you update regularly.

Payouts and bonuses

best payout online casino uk

8 queries · 4,340 searches/mo

Qualifying criteria — dedicated sub-pages.

We deliberately do not publish specific thresholds, signal weights or internal formulas: tuning them is exactly what separates a usable site structure from mechanical word grouping.

Privacy and data protection

The tool is built so that a confidential keyword set can be processed without risk to your client or project.

Processed in your browser

All algorithmic methods run locally on your device. Keyword lists are never sent to a server and are never stored by us.

No training on your data

Premium-mode requests are processed once on the server and are never used to train models or shared with third parties.

Credentials stay isolated

Access to external services lives only in encrypted server-side storage. Nothing is present in page code or the browser — no key can be extracted from the frontend.

History under your control

Saved projects belong to your account, are visible only to you and are deleted on request at any time.

Algorithms

All clustering methods — when to use each

Every algorithm has its own strength. Below is what type of keyword set each method suits, the volume it handles best, and the result you get.

Auto (Smart Recommendation)

When to use it
You don't want to pick manually — the system profiles your keyword set.
Volume
any volume
Result
An automatically selected method for your data, with an explanation.

Google SEO profile

When to use it
Classic search marketing: informational + commercial queries.
Volume
up to 50,000 keywords
Result
Groups ready to map onto site pages and intent types.

iGaming profile

When to use it
Narrow niche with brands, providers, bonuses and geo modifiers.
Volume
up to 50,000 keywords
Result
Clusters that respect brands, geo and typical niche patterns.

Shared-word overlap

When to use it
Uniform commercial sets with repeating phrasing.
Volume
up to 50,000 queries
Result
Fast, clear groups by shared words — the base skeleton of your structure.

Rare-word weighting

When to use it
Large sets where rare words must carry more weight.
Volume
up to 50,000 queries
Result
Groups where rare modifiers matter more than common words.

Word-fragment matching

When to use it
Sets with typos, different endings and word forms.
Volume
up to 30,000 queries
Result
Morphology tolerance: different forms of a word stay in one cluster.

Spelling-variant matching

When to use it
Lists with duplicates, misspellings and brand spelling variants.
Volume
up to 10,000 queries
Result
Near-identical queries merged and duplicates surfaced.

Topic hierarchy

When to use it
You need a tree of topics and subtopics, not a flat list.
Volume
up to 20,000 queries
Result
A cluster hierarchy that maps directly to site section levels.

Topic density

When to use it
Sets with dense topics and a lot of noise.
Volume
up to 20,000 queries
Result
Natural dense clusters plus outlier queries kept separate.

Query structure

When to use it
You need to split query types rather than topics.
Volume
no method limit
Result
Groups by pattern: "buy", "price", "reviews", "how to", geo queries.

Query length

When to use it
Quick segmentation before the main work.
Volume
no method limit
Result
Short head terms separated from long-tail queries.

Compare all methods at once

One run, several algorithms on the same set. Results appear side by side with cluster counts and distribution, so the choice is evidence-based rather than a guess.

Smart Recommendation

The system profiles your set — size, average query length, repetition, presence of volume data — and suggests the method that performs best on that specific data.

Search volume built in

A universal parser recognises volume columns in any format and sorts clusters by total demand instead of alphabetically.

Content analysisPro and Agency

Content Analyzer: page score and a ready content brief

The analyzer breaks any URL down into 20 quality factors: heading structure, text volume and density, topical coverage, readability, phrasing originality, presence of facts and numbers, internal links, metadata, semantic completeness and signs of generated filler content.

The output is not an abstract number but a report written in a senior SEO analyst voice: what exactly is weak, why it affects rankings and what to do next. The same analysis produces a copywriter brief in Markdown — H2/H3 structure, required topics, terms and recommended length.

Agencies get a bulk mode: upload an XLSX or CSV list of pages, get a processing queue and one report where every site sits on its own tab, with the same analysis and brief you see inside the tool.

  • 20 quality factorsStructure, topical coverage, readability, factual density, metadata and semantics in a single score.
  • Automatic Markdown briefA ready article outline with topics, terms and target length you can hand to a writer.
  • Your interface languageAnalysis and briefs are generated in the language you work in, not only English.
  • Bulk modeXLSX or CSV in, processing queue, multi-tab report out — one tab per site.
  • Analysis historyEvery report is stored, so you can return to it and compare after the text is reworked.
Open Content AnalyzerSee plan limits
factors in every analysis
20factors in every analysis
analyses per day on Pro / Agency
100 / 500analyses per day on Pro / Agency
bulk report with tabs
XLSXbulk report with tabs
Competitor analysisPro and Agency

The Force: text analysis of the entire Google Top 10

The Force takes your query, pulls the real Google results for the region and language you need, downloads the text of every page in the Top 10 and breaks it down into terms. You see what ranking content is actually made of instead of guessing.

Each term gets a Term Relevance Score based on how many competitors use it and how often. That produces a clear list: what to add to your text, what is underused, and what is over-optimised.

SERP results are cached, so repeating the same topic returns instantly and does not consume your limits. The SERP source, depth and geo are configurable per project.

  • Term Relevance ScoreA quantitative importance score for every term instead of vague "add more keywords" advice.
  • What to add, what to cutYour page compared to the Top 10: missing terms and excessive repetition in one list.
  • Geo and languageSERP analysis for a specific country and language, matching the market you write for.
  • 48-hour SERP cacheRepeat runs on the same topic do not eat into your daily limits.
  • Fallback SERP sourcesA provider chain with fallbacks so analysis never stops because one API is unavailable.
Go to The ForceSee plan limits
competitor pages per analysis
10competitor pages per analysis
analyses per day on Pro / Agency
30 / 150analyses per day on Pro / Agency
SERP cache with no limit usage
48hSERP cache with no limit usage
SERP and market dataPro and Agency

SERP clustering and Ahrefs data in one place

The most accurate way to know whether two queries need separate pages is to look at the SERP. SERP clustering pulls the top results for every keyword and merges queries whose results overlap. That is the same logic Google itself applies, not a word-based approximation.

To make it work on tens of thousands of keywords rather than hundreds, the engine uses our own comparison optimisation: it never brute-forces every pair, so large sets do not stall processing. Cluster formation streams in real time, and partially collected data can be clustered without waiting for the whole queue.

Market metrics from Ahrefs — search volume, CPC and keyword difficulty — sit next to the clusters, with aggressive caching so you never pay twice for the same lookup. Native exports from the popular SEO services are detected automatically, with no manual file conversion.

  • Clustering by SERP overlapQueries merged by the real URLs ranking for them — as close to Google's own logic as it gets.
  • Scales past 10,000 keywordsOur own comparison optimisation: large sets process without a speed collapse.
  • Volume, CPC, KDAhrefs market metrics next to each cluster so you can prioritise by potential immediately.
  • Project historyRuns are saved: come back to a large job and continue where you left off.
  • Complete exportA CSV with every keyword, cluster and metric — nothing truncated.
Go to SERP clusteringSee plan limits
SERP keywords per month on Pro / Agency
2,000 / 15,000SERP keywords per month on Pro / Agency
Ahrefs keywords per month
5,000 / 25,000Ahrefs keywords per month
SERP cache to save your quota
14 daysSERP cache to save your quota
WorkflowAll plans

Export, project history and data privacy

The tool should fit your process instead of forcing manual data transfers. Any result exports to CSV, JSON, TSV — including a grouped view where each cluster is its own block, and multi-tab reports for client deliverables.

Project history depends on your plan: 7 days on Free, 90 days on Pro and unlimited on Agency. That lets you revisit old keyword sets, compare iterations and hand work over inside a team.

Basic clustering runs in your browser — the keyword set never leaves your device. Heavy computation moves to a Web Worker, so the interface stays responsive even on tens of thousands of queries. The interface is translated into 30+ languages, including right-to-left scripts.

  • Export formatsCSV, JSON, TSV, grouped view and clipboard copy.
  • Project history7 days on Free, 90 days on Pro, unlimited on Agency.
  • Private by defaultBasic clustering runs locally in the browser — your keyword set is not uploaded.
  • No freezingLarge volumes are computed in a Web Worker while the interface stays responsive.
  • 30+ interface languagesFull localisation with RTL support for Arabic and Hebrew.
Start for freeCompare plans
keywords per run on Agency
50,000keywords per run on Agency
export formats
6export formats
interface languages
30+interface languages

Benefits of clustering

Why thousands of SEO specialists use our tool daily

Increase organic traffic

Properly structured content ranks higher in Google. Clustering helps create pages with maximum relevance.

+47% traffic

Save time

Automatic processing of thousands of keywords in seconds instead of hours of manual work. More time for strategy, less for routine.

10x faster

Increase relevance

Each page focuses on one topic with all related keywords. Google better understands your content.

+35% CTR

Logical site structure

Clusters naturally form page hierarchy: categories, subcategories, articles. Improves UX and indexation.

Avoid cannibalization

One page = one cluster = one search intent. No conflicting pages fighting for the same keywords.

-90% duplicates

Scalability

Process up to 50,000 keywords per request. Perfect for large projects and agencies.

Keywords per request
50K+Keywords per request
Clustering algorithms
7Clustering algorithms
Average processing time
<3sAverage processing time
Free forever
100%Free forever

Who uses clustering?

Tool for professionals and businesses of any scale

SEO Specialists

Building semantic core, planning site structure, optimizing content strategy

Example: Clustering 10,000 keywords for an electronics online store

Digital Agencies

Fast processing of large keyword volumes for client projects

Example: Client site audit and restructuring in 1 day

Copywriters

Planning articles with full topic coverage and all LSI keywords

Example: Creating content plan for 100 blog articles

E-commerce

Optimizing product categories and creating SEO descriptions

Example: Structuring catalog of 5,000 products

Bloggers

Finding article topics and avoiding content duplication

Example: Niche analysis and publication planning for the year

Business Owners

Understanding what customers search for and optimizing site for their needs

Example: Competitor analysis and finding untapped niches

Pricing

What the Free, Pro and Agency plans include

Keyword clustering is free and needs no signup. Paid plans lift the limits and unlock AI semantic methods, Content Analyzer, The Force, SERP clustering and bulk reports.

Free

To get started

$0/mo

Core clustering, no signup required

  • Up to 500 keywords per run
  • 5 clusterings per day
  • 8 clustering algorithms
  • Export to CSV, JSON, TSV
  • 30+ languages supported
  • Cluster visualization
Start free
Recommended

Pro

For SEO specialists

$29/mo

The full toolkit for daily SEO work

  • Up to 15,000 keywords per run
  • 100 clusterings per day
  • AI semantic methods (20/day)
  • Content Analyzer: 100 analyses per day
  • The Force: Top-10 analysis (30/day)
  • SERP clustering: 2,000 keywords per month
  • Ahrefs Volume/CPC/KD: 5,000 keywords per month
  • Bulk XLSX and Markdown reports
  • API access: 1,000 calls per month
  • 90-day project history
  • Priority support
Choose Pro

Agency

For agencies and teams

$89/mo

High volumes and bulk reporting for client projects

  • Up to 50,000 keywords per run
  • 500 clusterings per day
  • AI semantic methods (100/day)
  • Content Analyzer: 500 analyses per day
  • The Force: Top-10 analysis (150/day)
  • SERP clustering: 15,000 keywords per month
  • Ahrefs Volume/CPC/KD: 25,000 keywords per month
  • Bulk content analyzer
  • API access: 10,000 calls per month
  • Unlimited project history
  • Priority 24/7 support
Choose Agency

Enterprise

For large companies

Custom

Custom limits, SLA and integrations

  • Everything in Agency
  • Unlimited quotas
  • Dedicated capacity
  • SLA guarantees
  • Dedicated manager
  • Custom integrations and API
  • White-label option
  • Team training
Contact us
Detailed comparison of features and limits

Daily run limits reset every 24 hours; monthly SERP and Ahrefs volumes reset each month. Change or cancel your plan at any time.

100% privacy
Local processing
Instant results
Free forever

4.9 out of 5 stars

Based on reviews from 1,247 users

12M+
Keywords processed
8,500+
Active users
47
Countries worldwide
99.9%
Service uptime

🔒 Your data never leaves your browser. All processing happens locally.

Frequently Asked Questions

Everything you need to know about keyword clustering for SEO

Keyword clustering is the process of automatically grouping semantically similar search queries. One cluster = one relevant page on your site. This is the foundation for building site structure and content strategy.

Proper clustering can increase organic traffic by 30-50% through better content relevance.

Clustering helps: 1) Avoid keyword cannibalization when multiple pages compete for the same queries. 2) Create logical site structure. 3) Optimize internal linking. 4) Increase relevance of each page.

Optimally 5-20 keywords per cluster. Less than 5 — consider merging with another cluster. More than 20 — consider splitting into subclusters or creating a series of pages.

If a cluster contains more than 50 queries — this signals creating a hub page with subpages.

Auto — if unsure, system compares all methods. By Intent — for e-commerce and landing pages. By Rare Words — for niche sites. By Structure — for information portals. By Tail — for local SEO.

Similarity (0-100%) shows how well a keyword matches the cluster topic. >80% — perfect match, content foundation. 60-80% — good, include it. 40-60% — questionable, check manually. <40% — probably belongs to another cluster.

Filter results by similarity to exclude irrelevant queries.

Silhouette Score (-1 to 1) — clustering quality metric. >0.5 — excellent, clear groups. 0.25-0.5 — good, some overlap. 0-0.25 — satisfactory, consider another method. <0 — poor, clusters are mixed.

Available formats: CSV — for Excel, Google Sheets, Power BI. Contains cluster, keyword, similarity. JSON — for developers and CMS integrations. Structured format with metadata.

1) Export to CSV. 2) Each cluster = one article/page. 3) Cluster name — main keyword for Title. 4) Other cluster words — LSI for text. 5) Similarity determines inclusion priority.

Sort clusters by size — larger clusters usually have higher search potential.

No, absolutely not. All processing happens locally in your browser using JavaScript. Your data is not sent to server, not stored, and not analyzed by third parties.

Yes. Since processing is completely local, your business data including competitive analysis and semantic core remain confidential. No data leaves your browser.

Up to 50,000 keywords per request. Speed depends on your device. Approximately: 1,000 words — 1-2 sec, 10,000 — 10-20 sec, 50,000 — 1-3 min.

For large lists we recommend using By Intent method — it's the fastest.

Possible reasons: 1) Too many keywords (>10,000). 2) Using resource-intensive method (Semantic, Hierarchical). 3) Weak device or many open tabs. Try closing other programs.

Any format is supported: TAB-separated (keyword TAB 1000), semicolon (keyword;1000), in brackets (keyword (1000)), or space-separated (keyword 5000). You can also drag-and-drop a CSV/TXT file. The system automatically detects the format.

Numbers below 100 (e.g., 'iphone 15', 'top 10') are NOT parsed as frequency — they remain part of the keyword.

The system automatically sums up the search volume of all keywords in a cluster. This shows the total search potential of each group and helps prioritize content for the highest-volume clusters.

Sort clusters by total frequency to find the most valuable SEO topics.

Any exports from popular SEO tools are supported: Serpstat, Ahrefs, SEMrush, Google Keyword Planner, as well as CSV, TSV, TXT files. The system automatically finds the required columns (keyword, volume, etc.) regardless of their position in the file. Just drag-and-drop the file — everything else is handled automatically.

No need to edit files — the system smartly detects columns even if the file has lots of extra data.

Simply export keywords from any of these services in CSV or Excel format, then drag-and-drop the file into the input field. The system will automatically find columns with keywords and search volume, regardless of file structure and column names.

All report types are supported: positions, competitors, keywords, domain analysis, etc.

Algorithmic methods compare words and statistics, so synonyms and different phrasings of the same intent can land in different clusters. The AI mode evaluates search intent and topical proximity with a language model and adds specialised profiles for Google SEO and iGaming. It is available on Pro (20 runs per day) and Agency (100 runs per day), with up to 500 keywords per run.

Content Analyzer scores a single page of yours across 20 quality factors and generates a content brief. The Force works with the SERP: it pulls the Google Top 10 for your query, analyses the text of every competitor and calculates a Term Relevance Score for each term, showing what your page is missing. In practice they are used together — The Force supplies the terms, Content Analyzer validates the finished text.

Yes. Pro and Agency include a bulk mode: upload an XLSX or CSV list of URLs, the system queues them and produces one XLSX report where each site has its own tab with the analysis and the content brief. Exports to CSV, JSON, TSV and Markdown are also available.

Free gives you up to 500 keywords per request, 5 clustering runs per day, all 8 algorithmic methods, exports and 7 days of history — no signup required. Pro raises the limit to 15,000 keywords and 100 runs per day and unlocks AI methods, Content Analyzer, The Force, SERP clustering, Ahrefs data and bulk reports. Agency is built for teams: 50,000 keywords, 500 runs per day and unlimited project history.

Our algorithm is the grouping logic: how queries are normalised, how their closeness is measured, how clusters are formed and refined, and how the system picks a method for your keyword set. External services and data are sources of facts: search results for SERP methods, volume and difficulty metrics, and language models for semantic modes and brief generation. In short: facts come from outside, but the decision about what belongs together is made by our own logic.

That is why clustering stays reproducible even when external data changes: the inputs change, the grouping principle does not.

Algorithmic clustering runs directly in your browser — keyword lists never have to leave your device. For premium modes (AI, SERP, content analysis) the request is processed on our server and is never used to train models. All credentials for external services are stored server-side in encrypted storage, never appear in page code and are never visible in the browser. You can delete your project history at any time.

If your keyword set is confidential, use the algorithmic methods: the data stays on your device.

Total 22 questions for your convenience