Skip to content

text-statistics

How long is this text, how long will it take to read, and how hard is it? Count the words, sentences, and characters, average the lengths, divide by a reading speed, and blend the signals into a difficulty score — the numbers writing tools surface alongside every draft.

text-statistics is a pure-Rust, zero-dependency crate that computes all of it in a single pass: word / sentence / character counts, average word and sentence length, estimated reading time (seconds and minutes), and a 0–100 lexile-like difficulty estimate. It is the measurement engine behind the BeLikeNative AI writing assistant, packaged as a small library you can drop into editors, CLIs, and build pipelines.

What the numbers mean

signal meaning
words whitespace/punctuation-delimited word tokens
sentences ., !, ? terminators (a trailing fragment still counts)
characters total Unicode scalar values, including whitespace
avg_word_length letters per word (digits excluded, so 12345 is neutral)
estimated_reading_* seconds/minutes at 200 words per minute
difficulty_score 0–100 blend of word length and sentence length signals

The difficulty bands derived from the score:

band score style
Easy 0–32.99 short words, short sentences
Medium 33–65.99 typical adult prose
Hard 66–100 long words or long sentences

Install

[dependencies]
text-statistics = "0.1"

Quick start

use text_statistics::{stats, Difficulty};

let s = stats("The cat sat on the mat. It was a good day for a nap.");

println!("words:     {}", s.words);                    // 14
println!("sentences: {}", s.sentences);                // 2
println!("chars:     {}", s.characters);
println!("avg word:  {:.1}", s.avg_word_length);
println!("read sec:  {:.0}", s.estimated_reading_seconds);
println!("difficulty:{:.1}", s.difficulty_score);
println!("band:      {:?}", s.difficulty_band());      // Easy

See it live

The crate gives you the numbers. For a live in-editor word count, reading time, and difficulty readout with rewrite suggestions, use the BeLikeNative writing assistant.

Features

  • Pure Rust, no deps — a single file of arithmetic over the input, no features.
  • Single-pass totals — counts, averages, reading time, and difficulty in one walk.
  • Contractions kept whole — don't is one word, not two.
  • Digits don't inflate length — average word length counts letters only.
  • Difficulty bands — Difficulty enum + difficulty_band() for quick labels.
  • Deterministic — same input always yields identical output.
  • MIT licensed — embed anywhere.