A typing speed test answers two numbers and stops there: how fast the matching characters arrived, and what share of what you typed was right. This typing test uses one short passage. The clock starts on the first change in the box and stops when the box matches the passage exactly. Words per minute, the WPM figure, is not a count of dictionary words. For English and for Bahasa Malaysia the page uses the classroom rule that five characters make one word. Spaces count. A long word is several units, and a short word is less than one. That is why two people who "typed the same sentence" can land on different WPM if one of them hit extra letters and then fixed them: only correct characters enter the speed, and every typed character enters the accuracy.
The formula is small enough to check. Take the correct characters, divide by five for English or Malay, and divide by the minutes elapsed. One minute with 200 correct characters is 40 WPM. Thirty seconds with the same 200 correct characters is 80 WPM, because half a minute is half the divisor. The page will not divide by zero: a finished attempt always has at least a millisecond on the clock. Accuracy is correct divided by typed, times 100, then rounded. If you typed 100 characters and 97 matched the passage at those positions, accuracy is 97 percent. A character is compared one by one in order, including spaces. The test does not skip ahead to find a later match, and it does not ignore a double letter.
Chinese is scored on a different unit because the five-character word would invent a boundary the passage does not have. On the Chinese line, one correct character is one unit. A finished line of 30 characters in one minute is 30 on the same WPM label. That number is not comparable with an English 30 WPM, and the page does not pretend it is. The Chinese passage is a single Simplified Chinese sentence about a river, a bridge, a market, and a bus. The English passage is the same scene in English. The Malay passage is the same scene in Bahasa Malaysia, not a word-for-word gloss with English left in the middle of the sentence. The menu follows the language of the page when it loads, then you can switch. Switching clears the box and starts a new attempt.
Read the passage before you type. The box is a plain textarea with spellcheck off, so the browser does not underline a word and tempt a click that would leave the line. Type until the line is green and the result appears. A red character is a mismatch at that position. Backspace and fix it. The result shows WPM and accuracy, and it writes the latest score into this browser only. Come back later and the previous result is labelled as previous, until you finish a new line. Clearing the site's data for this host removes it. There is no leaderboard and no name field.
What a "good" WPM is depends on the keyboard, the language, and whether you are copying or composing. Many people who type every day land somewhere around 40 to 70 WPM on an English copy test of this length. A number far above that on a one-line passage can be a short clock rather than a lasting speed: the first keystroke starts the timer, so a pause before you begin does not count, and a very short line rewards a burst. Use the score to compare your own attempts on the same passage. Do not treat it as a certificate, a medical measure of dexterity, or the score a recruiter's software would print. Those tests set their own passage length and their own penalty for corrections.
If a key does nothing, the keyboard tester shows which key the browser received, including rollover when several keys are down. This typing test does not draw a keymap. If you wanted styled letters rather than a speed, the fancy text generator copies Unicode shapes and does not time you. The tool directory lists the rest. Nothing you type here is uploaded. The passage is fixed, the arithmetic is in the page, and closing the tab ends the attempt except for that one stored score.