Count words, phrases or survey lines
Use Words for ordinary prose. The counter groups repeated Latin-script words and can merge upper and lowercase spellings. Use Whole lines as phrases when each line is a complete name, expression or reviewed theme. It counts exact trimmed lines; it does not discover phrases inside sentences.
Every occurrence gives a word one count each time it appears. Once per non-empty line gives that word at most one count on each line. For a survey, put one complete response on a single line before using line counts. A line is an input boundary, not proof of a unique person. See a worked survey counting example.
Try a small example you can verify
Friendly friendly helpful
Friendly clear
Not clearSelect Try survey example to load these three invented responses with stop-word removal turned off. Friendly has three occurrences across two lines; clear has two occurrences across two lines; helpful and not each appear once. Switch the weight unit and count again to see the difference in the CSV weight column.
Review what your settings remove
The common English list is optional and includes words such as not and no. Your additional exclusions match complete entries. Minimum count and minimum character length hide entries after counting. The result separates parser exclusions from these minimum filters, so you can review what changed rather than assume every source word survived. Choose stop words without losing meaning.
Download counts and continue to artwork
Download weighted CSV uses the two-column word,weight format accepted by this site. Download text report includes both occurrence and line counts, selected settings, denominators and excluded entries. The table shows up to 250 entries; downloads include every retained entry, up to the 1,000-entry limit. Import that CSV in the weighted generator, review the list, and generate the layout.
Shares use all counted occurrences before minimum filters, or all non-empty input lines when line weights are selected. Line shares can add up to more than 100 percent because one line may mention several words. Word cloud font size is a design mapping of weights; use the count table when exact values matter. Understand how weights map to font size.
Know the counting limits
Text stays in this tab and is not automatically saved. The counter accepts at most 100,000 Unicode code points, 1,000 distinct retained parser entries and UTF-8 TXT files up to 512 KiB. Word entries longer than 64 characters and whole-line phrases longer than 100 are reported as excluded. Non-Latin scripts are not segmented in word mode. No stemming, synonym merging or sentiment classification is performed. Read the exact counting method and edge cases.