Skip to content
keysprint

About

A typing test that does the measurement and gets out of the way

keysprint measures how fast you type and how accurately, then explains what it measured. There is no account, no advertisement, no upsell and no streak to maintain. The test runs in your browser, the passages are written for the site, and your score never leaves your device.

What keysprint is

keysprint is a typing test with a practice programme attached. The test itself is the centre of the site: a live passage, a countdown, and a results panel that reports gross speed, net speed, three separate accuracy figures and an error count. Everything else, theguide, thelessons and thepractice drills, exists to give that number something to act on.

It is a small site on purpose. There is no leaderboard, because a leaderboard rewards the people with fast keyboards and long practice sessions rather than the people who want to improve. There is no account, because an account means a database, and a database means your data sits somewhere. There are no notifications, no streaks and no daily goals, because none of those things make anyone type faster.

What is left is the part that helps: a clear measurement, an honest explanation of how it was calculated, and specific exercises for the specific weaknesses the measurement reveals.

How the numbers are calculated

Every figure on the results panel comes from one of these five formulas. Nothing is rounded in your favour.

FigureFormulaWhat it means
Gross speed(characters ÷ 5) ÷ minutesHow many characters your fingers produced, counted as five character words per minute.
Net speed((characters − errors) ÷ 5) ÷ minutesThe same calculation after removing every character that does not match the passage.
Character accuracy(characters − errors) ÷ charactersThe share of the text on screen that is correct right now.
Word accuracycorrect words ÷ words typedThe share of whole words that match from the first letter to the last.
Keystroke accuracycorrect key presses ÷ key pressesThe share of raw key presses that were right the first time, including mistakes you later fixed.

One word is five characters, which is the convention that makes a score comparable between languages. The timer starts on your first keystroke rather than when the page loads, so reading the passage before you begin is not counted against you. And the clock stops on the deadline rather than when the next frame is drawn, so a tab you switch away from cannot quietly add time to your score.

What happens to your data

Your typing results stay on your device. The typing test runs entirely in your browser: the passage, the clock, the scoring and the results are all computed on your own machine. The site does use Google Analytics to measure page traffic, but it never receives your keystrokes or your scores.

  • No accounts. There is no sign-up, no login and no profile.
  • No advertising. There are no ads, affiliate links or advertising pixels.
  • Analytics for page traffic. Google Analytics 4 may set cookies and receive technical information such as your approximate location and browser type. It does not receive what you type.
  • Local storage for results. Your best score for each test length and your light or dark theme preference stay in your own browser.
  • No email addresses. You can use the whole site without giving a contact address.

The full details are in the privacy policy.

How it is built

The site is a static build. There is no application server, no database and no backend of any kind, which is the technical reason the privacy policy is so short. Every page is pre-rendered HTML, and the typing test is a small piece of client-side JavaScript that measures what you type and nothing else.

The scoring code is kept in one module with no browser dependencies, which means it can be unit tested against a fake clock. Those tests run as part of every build, so a change to the maths cannot ship without the checks going green.

The design is a single accent colour on a neutral scale, with light and dark themes that follow your system preference until you choose otherwise. It respects your reduced motion setting, works with a keyboard alone, and is legible at any zoom level.

Who it is for

  • People learning to touch type. Start with thehome row lesson, then work outward through thelesson stages.
  • People who type for a living. A weekly one minute test is enough to catch a regression, and theaccuracy test is the one to watch if your speed is fine but your corrections are not.
  • People who just want a number. Themain test pagegives you a score in about a minute, with nothing to sign up for.
  • Teachers and managers. The site works on a shared computer and needs no installation, which makes it convenient for a classroom or a workshop.

Getting in touch

Corrections, suggestions and bug reports are all welcome, and thecontact page explains how to send them.

Frequently asked questions

Questions about keysprint itself.

Who made keysprint?
keysprint is an independent, small project built by one person, with no company behind it, no investors and no marketing budget. It exists because most typing test websites either gate the useful parts behind an account or bury a simple measurement under enough advertising to be unpleasant.
Is keysprint really free?
Completely. There is no premium tier, no trial period and no paid feature that unlocks the results. The site runs on a free hosting tier and the test runs entirely in your browser.
Do you have accounts, leaderboards or progress tracking across devices?
No. Your best score for each test length is kept in your own browser using local storage, so it never leaves your device and never follows you to another one. Adding accounts would mean storing personal data, and keysprint is built specifically so that it does not have to.
Is the typing text original?
Every passage in the corpus was written for keysprint. Nothing is scraped from a novel, a news site or another typing test. The material is deliberately plain so it exercises the keyboard rather than the reader, and it is licensed for use here because it was written here.
How accurate are the measurements?
The scoring maths is defined in one place in the code, explained on this page, and covered by unit tests that run on every build. The one honest limitation is that a short test is noisier than a long one, which is why the page for each length explains what it is good for.