What Keyword Research Tool does
The Keyword Research Tool performs page-based term discovery rather than market-wide keyword research. It fetches one URL, extracts visible text, tokenizes up to a bounded amount of content, removes a small English stop-word list, and returns the twenty-five most frequent entries for each phrase length.
The output is useful for auditing what a page actually says, finding recurring terminology, and comparing language across drafts. It does not query a search index or advertising platform, so it cannot tell you monthly volume, search intent, cost per click, competitiveness, or where the page ranks.
How to use Keyword Research Tool
- Enter a public page URL that contains the content you want to study.
- Select Research page to fetch and analyze its visible HTML text.
- Review total word count plus the one-, two-, and three-word phrase lists.
- Use the findings with audience research and editorial judgment before revising the page.
Important details
This is a page corpus, not a search corpus
Every phrase comes from the submitted page's extracted text. The tool does not retrieve related queries, competitor pages, search results, ad-planner data, or user behavior.
Counts and densities are transparent
Each result includes how often the phrase appears and its percentage of usable phrase windows for that n-gram size. Alphabetical order breaks ties between equal counts.
Language handling is intentionally simple
Unicode letters and numbers are recognized, but the stop-word list is small and English-specific. There is no stemming, lemmatization, synonym grouping, or language-specific morphology analysis.
Practical uses
- Create a terminology inventory before consolidating overlapping web copy.
- Check whether a page visibly supports the topic promised by its title and description.
- Compare recurring phrases before and after an editorial rewrite.
Privacy and limitations
The submitted URL goes to PagesTools, and the server requests the public page for analysis. The destination can record the server request; the resulting JSON report uses no-store cache headers.
The tool analyzes one bounded HTML response without executing JavaScript. It does not supply search demand, trends, ranking positions, competitors, intent labels, or recommendations about keyword targets.