Corpus analysis for Node.js

Read between
the words.

Turn raw text into reliable signals. Exact word frequencies, recurring phrases, lexical statistics, and language-aware views— all in one focused library.

Try the analyzer View on npm

01 / Frequency map sample.txt

language signal meaning patterns text exact context words

01040812

ESM CommonJS Node 20+

Small API. Auditable results. No black box.

Word frequencies◆ N-grams◆ Concordances◆ Lexical statistics◆ 7 language profiles◆

[ 01 ]

One corpus, many views

See what your text
is really saying.

01 Core analysis

Count with confidence.

Stable rankings, exact maps, and Unicode-aware normalization. Same input, same answer—every time.

02 Language aware

Speak the language.

Built-in profiles respect the alphabet, locale, stop words, and stemming rules of your corpus.

03 Phrase discovery

Find phrases, not just words.

Count bigrams and trigrams across complete strings, incremental chunks, or readable streams.

04 Made for code

Fits your pipeline.

TypeScript declarations, ESM and CommonJS exports, immutable result views, and native stream support.

[ 02 ]

Live playground

Paste text.
Find the signal.

Input / corpus.txt

Profile Remove stop words

Output / frequency Complete

—Tokens

—Types

—Diversity

TermFrequencyCount

This browser demo mirrors core counting behavior. The package adds streams, concordances, stemming, n-grams, and exact result views.

Zero to insight in three lines

Small surface.
Serious analysis.

analysis.mjs

import { analyze } from "freqwords";

const result = analyze(feedback, {
  profile: "english",
  stopWords: true,
  ngramSizes: [2, 3],
});

console.log(result.entries("filtered", 10));

MIT licensed TypeScript ready Node.js 20+

Read between the words.

See what your text is really saying.