Lexical diversity analyzer
Measure transparent vocabulary ratios and moving-window variation in your own passage.
Word tokens are Intl.Segmenter word-like segments containing a Unicode letter; pure numbers are excluded. TTR = distinct words / words; root TTR = distinct words / square root of words; corrected TTR = distinct words / square root of twice the words. Moving-average TTR uses only complete selected-size windows. These length-sensitive descriptive ratios do not measure proficiency or writing quality. Use up to 500000 text characters and 20000 word tokens. Reports allow up to 10000 matches or repeated phrase groups. JSON and CSV preserve the full report, with bounded exports up to 40 million characters and 500000 CSV records. Analysis runs locally with a 10-second timeout. Preview search and sort cover the first 200 rows; copy returns the summary. Download the full report for every row. Your text is not uploaded. Up to 500,000 characters. Review the settings before running.
Your result
How to use Lexical diversity analyzer
- Paste your own passage as plain text.
- Choose explicit matching or analysis settings.
- Review the bounded table and download complete JSON or CSV.
Useful next steps
Know exactly what is measured
Word tokens are Intl.Segmenter word-like segments containing a Unicode letter; pure numbers are excluded. TTR = distinct words / words; root TTR = distinct words / square root of words; corrected TTR = distinct words / square root of twice the words. Moving-average TTR uses only complete selected-size windows. These length-sensitive descriptive ratios do not measure proficiency or writing quality.
Keep the full report
Use up to 500000 text characters and 20000 word tokens. Reports allow up to 10000 matches or repeated phrase groups. JSON and CSV preserve the full report, with bounded exports up to 40 million characters and 500000 CSV records. Analysis runs locally with a 10-second timeout. Preview search and sort cover the first 200 rows; copy returns the summary. Download the full report for every row. Your text is not uploaded. Offsets refer to the original UTF-16 text. Machine report keys and your content are preserved. Formula-protected CSV changes risky cell prefixes; import columns as text when preserving exact identifiers.