AnyTool
Your files never leave your device. All processing happens locally in your browser.

Search Intent Classifier

Classify a keyword list by search intent (informational, commercial, transactional, navigational) from its signal words, with matched signals & a suggested content type — 100% client-side

Related Tools

All SEO & Web →
FAQ Gap FinderAMP ValidatorCanonical Tag CheckerHTML Sitemap GeneratorHTTP Header CheckerMeta Tag Analyzer

How do I classify a list of keywords by search intent — informational, commercial, transactional or navigational — and know what type of content to build, without uploading my keywords?

Open AnyTool’s Search Intent Classifier, paste your keywords one per line, and it labels each one entirely in your browser. It matches every keyword against curated signal-word (modifier) lexicons for the four search intents — informational (how, what, why, guide, tutorial, examples, meaning) → a blog post or how-to; commercial investigation (best, top, review, vs, comparison, alternatives, cheap) → a comparison or review; transactional (buy, price, coupon, deal, order, for sale, discount, subscribe, download, free trial) → a product or pricing page; and navigational (a brand name, login, official site, near me, or a URL) → a brand homepage or local page. For each keyword it shows the winning intent with a colour badge, a confidence, the exact signals it matched, an optional secondary intent for mixed-intent keywords (like “best running shoes to buy”, which is both commercial and transactional), and a suggested content type. You get a live breakdown with a distribution bar, an intent filter, and copy/CSV export. Everything is computed locally — your keyword list is never uploaded, so it is safe for confidential research. Honest caveat: this reads intent from keyword signal words, not from the SERP — it is a fast heuristic to triage a list, not a search-results analysis. The true intent is whatever Google actually ranks for the query, so always confirm the important ones by checking the live SERP; many keywords are ambiguous, and “no signal words” does not prove informational.

  • Classifies keywords into 4 intents (informational, commercial, transactional, navigational) from signal-word lexicons
  • Shows the exact signals matched, a confidence, a secondary intent for mixed-intent keywords, and a suggested content type
  • Bulk paste, live intent breakdown with a distribution bar, filter by intent, copy report + CSV export
  • 100% client-side — your keyword list is never uploaded; safe for confidential research
  • Honest: signal-word heuristic, not a SERP analysis — confirm the true intent on the live search results

What is

Search intent (keyword intent)

Search intent (also called keyword or user intent) is the goal behind a search query — what the person actually wants to accomplish. SEO practice groups it into four main types: informational (the searcher wants to learn something, e.g. “how does SEO work”), navigational (they want a specific site or place, e.g. “canva login” or “coffee shops near me”), commercial investigation (they are researching options before buying, e.g. “best noise-cancelling headphones” or “paid SEO suites vs enterprise SEO platforms”), and transactional (they are ready to act — buy, subscribe, download, e.g. “buy standing desk online”). Matching a page’s content type to the query’s intent is one of the most important ranking factors: a product page will struggle to rank for an informational query and vice-versa. Intent can be estimated from signal words (modifiers such as “how”, “best”, “buy”, “login”) as a fast triage, but it is most reliably confirmed by analysing the SERP — the type of pages Google already ranks (blogs, comparisons, product pages, brand sites) reveals the intent Google has decided the query has.

SEO & Web

Related terms

search intentkeyword intentuser intentinformational intentnavigational intentcommercial intenttransactional intentbuyer intent keywordsSERP analysiscontent typekeyword research

Frequently Asked Questions

Informational (the searcher wants to learn), navigational (they want a specific site or place), commercial (researching options before buying), and transactional (ready to buy, subscribe or download). Each maps to a different content type.

Search intent is the goal behind a query, and SEO practice groups it into four main types. Informational intent means the person wants to learn or understand something — queries with words like how, what, why, guide or tutorial — best served by a blog post or how-to. Navigational intent means they want to reach a specific website, page or place — a brand name, login, official site, near me, or a typed URL — served by that brand’s homepage, login or local page. Commercial (or commercial-investigation) intent means they are researching before a purchase — best, top, review, vs, comparison, alternatives, cheap — served by a comparison, review or roundup. Transactional intent means they are ready to act — buy, price, coupon, deal, order, subscribe, download — served by a product, pricing or landing page. Matching your page’s content type to the query’s intent is one of the most common reasons a page ranks or fails to rank, which is why classifying keywords by intent is a core keyword-research step. AnyTool’s classifier labels each keyword by these four types from its signal words and suggests the content type to build.

It is a fast, useful heuristic, not a definitive answer. Signal words like buy, best or how strongly hint at intent, but the true intent is whatever Google actually ranks for the query, so you should confirm important keywords by checking the live SERP.

Classifying intent from the words in a keyword — the modifiers like how, best, buy or login — is a fast, transparent way to triage a large keyword list, and for clear cases it is quite reliable: “how to…” is almost always informational and “buy… online” is almost always transactional. But it is a heuristic, not a definitive analysis. The most reliable way to determine intent is to analyse the search results themselves: Google has already evaluated thousands of pages to decide what satisfies a query, so the top results reveal the intent — if they are blog posts the intent is informational, if they are comparison pages it is commercial, if they are product pages it is transactional, and if they are one brand’s own pages it is navigational. Many keywords are also ambiguous or mixed-intent — “best running shoes to buy” is both commercial and transactional, and a bare word like “coffee” could be almost anything. So use a signal-word classifier to sort and prioritise quickly, then confirm the keywords that matter by checking the real SERP before you commit content.

Informational keywords call for a blog post, guide or how-to; commercial keywords for a comparison, review or roundup; transactional keywords for a product, pricing or landing page; and navigational keywords for the relevant brand homepage, login or local page.

Each search intent is best matched by a specific type of page. For informational keywords — how-tos, definitions, guides — create a blog post, tutorial or in-depth guide that teaches the topic. For commercial-investigation keywords — best, top, review, vs, alternatives — create a comparison, a review or a “best X” roundup that helps the reader choose. For transactional keywords — buy, price, order, subscribe, download — create a product page, pricing page or a conversion-focused landing page that lets the person act. For navigational keywords — a brand name, login, official site, near me — the right result is usually that brand’s own homepage, login page or local/store page, so this is mostly about your own branded terms and local listings rather than new content. The content format and angle matter too, which is why you confirm on the SERP: study what already ranks (a listicle, a step-by-step guide, a product grid) and match it. AnyTool’s classifier suggests a content type for every keyword so you can plan your content map by intent.

No. Your keywords are matched against built-in signal-word lists entirely in your browser with plain JavaScript and are never uploaded, logged or stored, so it is safe for confidential keyword research. It also works offline once cached.

The Search Intent Classifier is 100% client-side: the keywords you paste are matched against built-in signal-word lexicons with ordinary JavaScript on your device — nothing is fetched, uploaded, logged or stored on a server, and no CDN is called for processing. That makes it safe for unpublished keyword research, client lists and confidential content strategy without exposing them. It classifies live as you type, ignores duplicate lines, shows a live intent breakdown with a distribution bar, lets you filter the table by intent, and offers a copyable report and CSV export; it works offline once cached as a PWA, supports dark mode and a mobile layout with a sticky Sample / Copy bar. The logic is a small pure engine (searchIntentEngine.ts) with no new dependency — everything runs locally in your browser.

Detailed Explanation

How It Works

What the Search Intent Classifier Does

Search Intent Classifier is a 100% client-side tool that labels a list of keywords by search intent — the goal behind each query — live as you type. It uses the standard four-type taxonomy: informational (the searcher wants to learn something, e.g. “how to compress a pdf”), commercial investigation (researching before a purchase, e.g. “best noise-cancelling headphones” or “paid SEO suites vs enterprise SEO platforms”), transactional (ready to act, e.g. “buy standing desk online”), and navigational (wanting a specific site or place, e.g. “canva login” or “coffee shops near me”). Each keyword is matched against curated signal-word (modifier) lexicons — how / what / guide for informational; best / top / review / vs for commercial; buy / price / coupon / order for transactional; login / official site / near me / a URL for navigational — and the tool returns the winning intent with a colour badge, a confidence, the exact signals it matched, an optional secondary intent for mixed-intent keywords, and a suggested content type (blog vs comparison vs product page vs brand homepage). A summary shows counts per intent with a distribution bar, the table can be filtered by intent, and results copy or export to CSV.

  • Classifies keywords into 4 intents: informational, commercial, transactional, navigational
  • Matches curated signal-word lexicons (how/guide, best/vs, buy/price, login/near me)
  • Per keyword: intent badge, confidence, exact matched signals, optional secondary intent, suggested content type
  • Live intent breakdown with a distribution bar; filter the table by intent
  • Bulk paste (duplicates ignored); copy report and CSV export
Methodology

How Intent Is Scored From Signal Words In-Browser

All logic lives in a small pure module (searchIntentEngine.ts) with no new dependency and no DOM access. Each keyword is lower-cased and whitespace-normalised, then matched against four signal-word lexicons using word-boundary-aware phrase matching, so “how to” matches as a phrase and “buy” matches only as a whole word. Each matched signal adds to that intent’s score (multi-word phrases weigh slightly more than single words), and a few precedence rules refine it: a leading question word (how, what, why, where, who, which) gives a strong informational boost; a domain/URL token in the query (e.g. “nike.com”) gives a strong navigational boost; and when a keyword carries both commercial and transactional signals (e.g. “best running shoes to buy”) the action nudges it toward transactional while keeping commercial as the secondary intent. The highest-scoring intent wins, ties break toward the more specific intent, and confidence is derived from the winning score and its margin over the runner-up (high, medium or low). If a keyword matches no signals at all, it is defaulted to informational at low confidence and flagged — explicitly a starting guess, not evidence. The suggested content type is a fixed mapping per intent (blog/guide, comparison/review, product/pricing, brand homepage/local).

  • One pure engine (searchIntentEngine.ts), no new dependency, no DOM access
  • Word-boundary phrase matching against four curated signal-word lexicons; phrases weigh more than single words
  • Rules: leading question word → informational; a URL/domain → navigational; commercial+transactional overlap → transactional, commercial kept as secondary
  • Confidence from the winning score and its margin over the runner-up (high/medium/low)
  • No signal words → defaulted to informational at low confidence and flagged as a guess
Limitations

A Signal-Word Heuristic — Not a SERP Analysis

The tool is deliberately honest about scope. It reads intent from the modifier words in each keyword, which is a fast, transparent way to triage a large keyword list — but it is a heuristic, not a search-results analysis. The most reliable way to determine intent is to analyse the SERP: Google has already evaluated thousands of pages to decide what satisfies a query, so the type of pages that rank (blog posts, comparison pages, product pages, or one brand’s own site) reveals the true intent. So the true intent is whatever Google actually ranks for the query, and important keywords should always be confirmed on the live SERP before you commit content. Many keywords are also ambiguous or mixed-intent — “best running shoes to buy” is both commercial and transactional, and a bare word like “coffee” could be informational, commercial or navigational — which is why the tool surfaces a secondary intent where it sees one. And crucially, “no signal words” does not prove informational: those keywords are defaulted to informational at low confidence only as a starting point. The lexicons are curated but not exhaustive, and a modifier can be misleading in context. Use the labels to sort and prioritise quickly, then verify the ones that matter on the real search results.

  • Reads intent from keyword signal words — triage, not a SERP analysis
  • The true intent is whatever Google ranks for the query; confirm important keywords on the live SERP
  • Many keywords are ambiguous or mixed-intent — a secondary intent is surfaced where present
  • No signal words does NOT prove informational — such keywords are a low-confidence default
  • Lexicons are curated but not exhaustive; a modifier can mislead in context
Privacy & Security

Why Classifying Search Intent In-Browser Is a Privacy Win

Many keyword-intent tools send your list to a server or gate results behind an account or a paid plan. This tool instead treats your keywords as plain strings that are matched against built-in signal-word lexicons with ordinary JavaScript entirely on your device; nothing is fetched, uploaded, logged or stored, and no CDN is called for processing. That makes it safe to classify unpublished keyword research, client lists and confidential content strategy without exposing them. It classifies live as you type, ignores duplicate lines, shows a live intent breakdown with a distribution bar, lets you filter the table by intent, and offers a copyable report and a CSV export for your content map. It works offline once cached as a PWA, supports dark mode and a mobile layout with a sticky Sample / Copy bar, and has exactly three ad slots. The logic is a small pure engine (searchIntentEngine.ts) with no new dependency — so the classification is transparent (you can see exactly which signal words drove each label) and entirely local.

  • Keywords are matched locally against built-in lexicons — never uploaded, no server, no CDN for processing
  • Safe for unpublished keyword research, client lists and confidential strategy; no account or paywall
  • Live intent breakdown + distribution bar, filter by intent, copy report and CSV export
  • Works offline once cached; dark mode and mobile with a sticky Sample / Copy bar; exactly three ad slots
  • Pure engine (searchIntentEngine.ts), no new dependency — transparent, explainable labels
Search intent classification: in-browser (AnyTool) vs typical online tools
CapabilityAnyToolTypical online tools
ProcessingMatches signal words in your browserOften sends your keyword list to a server
Your keywordsNever uploaded, works offlineUploaded / processed server-side
Account / planNone — paste and goOften gated behind sign-up or a paid plan
Intents4 types + secondary for mixed-intent + suggested content typeOften a single label, no content-type hint
TransparencyShows the exact signals matchedUsually a black-box label
HonestyStates it is signal-based triage, not a SERP analysisOften implies definitive intent
CostFree, no sign-up, no limitsFree tier often limited / upsold

AnyTool is a 100% client-side search intent classifier: paste a keyword list and it labels each one by intent — informational, commercial, transactional or navigational — from curated signal-word (modifier) lexicons, showing the exact signals matched, a confidence, a secondary intent for mixed-intent keywords, and a suggested content type (blog vs comparison vs product page vs brand homepage), with a live breakdown, an intent filter and copy/CSV export. Honest caveat: this reads intent from keyword signal words, not from the SERP — it is a fast heuristic to triage a list, not a search-results analysis. The true intent is whatever Google actually ranks for the query, so confirm important keywords on the live SERP; many keywords are ambiguous or mixed-intent, and having no signal words does not prove informational. Comparison as of June 2026.