N-Gram Analyzer guide
Count and rank exact contiguous sequences of two to five words in a passage with transparent local processing.
What this tool does
The N-Gram Analyzer converts the entered text to lowercase word tokens and slides a fixed-size window across them. A two-word sequence is called a bigram, a three-word sequence is a trigram, and larger settings in this tool contain four or five words. Every exact sequence is counted and ranked by frequency, with alphabetical order used to break ties.
For the phrase local tools process local files, the two-word windows are local tools, tools process, process local, and local files. Word order matters: local tools and tools local are different n-grams. Punctuation between words is removed during tokenization, but the original sequence of words is retained.
This is an exact phrase-frequency utility, not a semantic model. It does not combine synonyms, stem related words, recognize quotations, infer topics, or determine whether repetition is helpful. A phrase that appears once can still be important even though repeated phrases rank above it.
How to use it
- Paste a complete passage into the editor.
- Choose a sequence size from two through five words.
- Choose how many ranked results to display.
- Select Analyze sequences.
- Review or copy the tab-separated phrase and occurrence table.
Use the same settings when comparing drafts. A larger sequence size is more specific but produces fewer repeats, especially in short text.
Benefits
- Counts contiguous phrases rather than isolated keywords
- Supports two-, three-, four-, and five-word windows
- Displays exact occurrence counts
- Copies results as a simple tab-separated table
- Keeps unpublished text on the device
Interpreting repeated sequences
N-grams can reveal recurring product language, repeated headings, boilerplate, accidental phrasing, translation consistency, and common support-message patterns. They can also help verify whether an exact multiword term appears consistently across a draft.
Frequent phrases are not automatically search keywords and should not be repeated mechanically. Search intent, usefulness, clarity, originality, and audience needs require broader evidence. Similarly, matching n-grams do not prove plagiarism or authorship because common terminology and standard expressions naturally recur.
Analyze licensed or authorized text, and review the original context before publishing conclusions based on counts.