Every figure follows a few plain rules, so you can predict the result before you read it. A line break (including Windows-style CRLF) counts as one character.
Words: a word is any run of non-space characters that contains at least one letter or number. Hyphenated words (well-known), contractions (don't), decimals (3.14) and plain numbers each count as one word. A stray dash, an ampersand or an emoji on its own is not a word. For Chinese, Japanese, Thai, Lao, Khmer and Myanmar text, which has no spaces between words, the tool asks your browser's word segmenter (Intl.Segmenter) where the words begin and end.
Characters: the count is of characters as a reader sees them, so an emoji, a flag, a letter with a combining accent, or a Thai or Lao consonant with its SARA AM vowel counts as 1. Characters without spaces leaves out every kind of whitespace: spaces, tabs, line breaks and non-breaking spaces.
Sentences: a sentence ends at one or more of . ! ? or an ellipsis, followed by a space or the end of the text. Consecutive marks such as ?! or ... count once, and 3.14 or example.com do not split a sentence. The full-width marks used in Chinese and Japanese, and the Hindi danda, end a sentence even without a following space. Text left over at the end counts as one more sentence.
Paragraphs: a paragraph is a block of text separated from the next by at least one blank line (a line that is empty or holds only spaces).
Reading and speaking time: both are estimates from the word count. The tool uses 238 words per minute for silent reading and 130 words per minute for speaking aloud, which are commonly quoted averages. Your own pace will differ, and the result is rounded up to the next whole second.