Modes and Exercises
In this typing speed test, two input language modes, English and Japanese kana practice, are provided to accommodate your training goals.
The English mode offers 6 predefined English sentences, while the Japanese mode presents 6 preset kana sentences accompanied by natural Japanese text displays.
You can easily switch between these two modes at any time from the settings panel located on the right side of the interface. The practice timer can be selected from three standard durations: 15 seconds for quick rhythm checks, 30 seconds for balanced pacing, and 60 seconds for extended concentration exercises.
This tool does not have a mechanism to dynamically fetch external text sources, save progress across sessions, or automatically generate new practice sentences using AI.
It sequentially reads fixed sentence arrays stored directly in browser memory. The Japanese mode draws upon a dedicated kana dictionary and preset tasks with authentic Japanese displays, although it operates as a structured romaji matching system rather than a universal, fully comprehensive IME converter.
Due to the fixed preset variations, taking multiple tests will cycle through the exact same collection of sentences. This makes the application particularly well suited for drilling specific key positions, romaji combinations, and fingering patterns to establish reliable muscle memory, rather than evaluating your spontaneous typing adaptability on unfamiliar text.
Because all keystroke detection and scoring happen locally within the client browser, you can immediately begin practicing and monitor your input cadence simply by typing on your keyboard.
Understanding Input Variations
The Japanese typing exercises actively support standard romaji variations, such as accepting "shi", "si", or "ci" for a single kana character, alongside compound syllable spellings like "tya" and "cha".
Rather than strictly restricting you to a single Latin spelling, the system references predefined dictionary aliases to evaluate your keystrokes, though it does not serve as a universal or exhaustive IME romaji conversion engine.
At the bottom of the screen, guide text indicating the next expected character is displayed in real-time to assist your typing flow.
The tool incorporates matching logic for special Japanese phonetic rules, including double consonants (sokuon) formed by repeating the initial consonant of the following syllable, as well as context-dependent "n" inputs.
While these rules provide solid support for standard kana typing, unlisted phonetic variations outside the dictionary are not recognized, meaning it does not guarantee complete coverage of every possible input method.
Contextual judgments are applied directly during typing, such as requiring "nn" before vowels or "y" while accepting a single "n" preceding most other consonants.
These evaluations occur on every keydown event through internal string matching functions that compare input characters against the valid romaji pathways.
While this design handles everyday transcription patterns comfortably, non-standard spellings or direct IME composition remain unsupported.
Starting and Switching Sentences
The measurement begins automatically to streamline the start of your exercise. After choosing your preferred duration (15, 30, or 60 seconds) and focusing the input field, typing starts the countdown.
A dedicated keyguard actively ignores key repeats, IME composition events, modifier keys such as Ctrl, Meta, or Alt, and multi-character non-text keys.
Furthermore, Backspace allows you to edit typed characters but will not trigger the timer; only the first accepted single-character keypress (even if it is a typo) initiates the test.
During active measurement, accurately completing the presented sentence causes the system to load the next preset sentence in the sequence, looping back to the beginning after the 6th sentence.
You can also manually skip to the next exercise at any moment by clicking the "write a different sentence" button located below the text.
When you advance to the next sentence—whether upon natural completion or via manual skipping—the application only clears your currently typed text while seamlessly continuing the running countdown and accumulating your ongoing statistics.
If you wish to start a completely fresh test with reset data, you must switch the language, change the test duration, or click the dedicated reset button.
When the designated time expires, the system detects the timeout and concludes the session. The countdown is calculated using a target deadline approach that evaluates remaining time via Math.ceil((deadline - Date.now()) / 1000) inside a 1000ms callback interval, compensating for timer drift rather than cumulatively subtracting seconds, although browser thread scheduling means absolute hardware precision is not guaranteed.
Once the test finishes, the input area locks to prevent additional keystrokes, and clicking the "reset" button restores the initial state so you can start over.
WPM and Accuracy
To objectively evaluate typing performance, this tool calculates two main metrics, "WPM (Words Per Minute)" and "Accuracy", and displays them in real-time on the screen.
WPM is a benchmark indicating the number of words that can be typed per minute, but to eliminate the unfairness of physical word lengths varying by sentence, this tool adopts the expedient formula of "considering 5 correct keystrokes as 1 word".
Specifically, it is calculated by round(correct/5/elapsedMinutes). The elapsed minutes are calculated from the remaining time in seconds, and to avoid zero division right after starting, a fallback of 0.01 minutes is used.
Therefore, the initial WPM fluctuates easily, but it allows you to grasp the relative input speed in 1-minute equivalents even in short measurement times like 15 or 30 seconds.
The decimal part is rounded after calculation, provided as an integer value.
Accuracy represents the percentage of correctly typed keys against the total number of key inputs made by the user. The formula is round((correct/total)*100), and this is also rounded to an integer for display.
Since the total keystrokes are zero immediately after starting, the system treats it as 100 for convenience if there is no input.
These score calculations are completed by updating state variables within the browser, rewriting the on-screen numbers instantly with each keystroke.
Incorrect Inputs and Reversion
If you press any key other than the specified character during a test, the application immediately detects it as a mistype. When an error takes place, the border of the active input field briefly flashes red to give you immediate visual feedback on your mistake.
Concurrently, the internal tracking logic increments the "error count" by 1 and increases the "total key inputs" counter without raising the "correct input" count, which causes your calculated accuracy percentage to drop accordingly.
The application includes a keyguard filter that actively ignores automatic key repeats, IME composition states, modifier keys like Ctrl, Meta, or Alt, and multi-character non-text keystrokes.
However, pasting text directly and traditional IME conversion remain unsupported, as the system evaluates individual single-character physical keydown inputs against expected characters.
Erroneous keystrokes are prevented from appearing in the visible input field, but each typo continues to accumulate in your error statistics.
Whenever you make a mistake, you can press the Backspace key to edit your progress and delete the previous character. Pressing Backspace moves the visual input position back by one character, but this action does not subtract from or undo the accumulated error count or total keystroke statistics.
Because previously registered correct keystrokes are not decremented upon deletion, re-typing those characters continues to increment the correct keystroke counter; consequently, frequent backspacing and re-typing can artificially inflate your calculated WPM score.
Once the test concludes, the results summary highlights a list of your top 5 difficult keys, determined by ranking the characters that were expected when mistypes occurred.
Repetitive Practice and Measurement Limits
All data processing and score judgment of this tool are performed solely on the browser's execution memory. The scorecard displayed after completion summarizes the final WPM, accuracy, total keystrokes, number of mistypes, and tally of difficult keys, but this data is never saved to a persistent area.
Reloading the page or closing the browser tab will result in the loss of all measurement results and statistical data. There is no saving, CSV export, or external fetching capability.
Therefore, it does not support long-term uses like saving and graphing daily growth records, but is designed for one-off repetitive practice or immediate warm-up purposes.
If you want to compare with past numbers or keep the results, you must take a screenshot of the screen or record the numbers directly in a notepad.
Furthermore, the WPM and accuracy calculated here are merely measurement results for the specific short sentences fixedly prepared in this tool.
It is not identical to the evaluation of long text creation in actual business, typing while thinking about the content, or the proprietary scoring criteria adopted by other typing sites.
In real tasks, various factors such as predictive conversion and parallel processing of thought and hands affect speed. Please use this tool as an indicator to measure "how fast you can physically trace the displayed short strings".