About SetupCheck

SetupCheck is a set of free tests for the hardware on your desk: camera, microphone, keyboard, mouse, gamepad, screen, speakers and headphones. They run in your browser, with no account and nothing to install, so you can find out quickly whether a device works as it should.

Who runs SetupCheck

SetupCheck is run by Maciej Barta WebForge. You can reach us at hello@setupcheck.app, and the contact page says what to write about.

How the tests measure

Each test uses a feature your browser already has: the camera and microphone interfaces, key and mouse events with timestamps, the gamepad interface, Web Audio for sound, and frame timing for the screen. A test reads what the browser reports and does the sums in the page, on your device. Where only your eyes or ears can judge, as with dead pixels or stereo channels, the test shows a pattern or plays a sound and asks what you saw or heard.

The thresholds a test uses, such as how much stick drift counts as drift, are in the code of that test, and the end of each test page explains them under “How we measure”.

Test by test

Camera

• Webcam test: counts the video frames that actually reach the page over the last two seconds, and takes the resolution and the reported frame rate from the settings the browser gives the camera. You judge sharpness and color by eye.

Audio

• Microphone test: reads the incoming sound and shows the peak and average level in dBFS, decibels below the loudest signal a digital system can carry. It flags clipping and a signal that is too quiet. dBFS is not a loudness measurement in a room.
• Speaker test: plays test sounds on each channel through Web Audio. The page does not listen back, so the verdict is built from your answers.
• Headphone test: plays channel, balance, bass and polarity checks and a frequency sweep from 20 Hz to 20 kHz, and you mark where you start to hear the sweep and where it fades. It is not a hearing test.
• Tone generator: makes a steady tone from 1 to 20,000 Hz with the sine, square, sawtooth or triangle waveform. The frequency is exact in software, while what comes out depends on your hardware.

Keyboard and mouse

• Keyboard test: reads key down and key up events for each physical key, with timestamps. A key pressed again within 30 ms of its release is flagged as possible chatter, and a held key shows its repeat delay and rate.
• Mouse button test: counts presses per button and times your double-clicks. A press that follows the previous one too closely to be deliberate is flagged as a possible double-click from a worn switch.
• Typing speed test: measures words per minute, counting five characters as one word, and accuracy as the share of correct keystrokes. The timer starts with your first letter.

Gamepad

• Gamepad test: reads the controller through the browser’s gamepad interface. It samples the sticks at rest for two seconds and reports drift when a stick rests more than 10% of its travel from the center.

Display

• Screen resolution: shows what the browser reports about your screen, window and pixel ratio. Some browsers round or hide values, and the browser may not know the system’s exact scaling mode (for example on a Mac).
• Refresh rate test: times the frames the browser draws and reports a rate when they arrive evenly. This is usually the screen’s rate, but power saving or a busy system can lower it.
• Ghosting test: moves shapes at fixed speeds over black, gray and white, and you judge any trails or halos. The page checks only that frames arrive on time and gives no response time figure.
• Dead pixel test: fills the screen with solid colors so stuck or dead pixels stand out. The page measures nothing, because it cannot see the panel.
• Backlight bleed test: shows dark shades so you can look for light leaking at the edges, glow and clouding. The summary uses only your answers.

Reflex

• Reaction time test: runs five rounds with a random wait of 1.5 to 5 seconds, and times from the moment the signal is drawn to your press. A press before the signal, or under 100 ms after it, counts as too soon. The result includes the delay of your screen and input device, so compare it only on the same setup.

What the tests cannot do

• They cannot see the hardware itself. A test sees what the browser and the operating system report, so a result describes the device in this browser, with its drivers and settings at that moment.
• Other apps, power saving and privacy settings can change a result. Some browsers round their clocks, which switches off the timing checks that need a fine clock.
• The tests are not calibrated or certified, and they do not repair or diagnose. A good result does not prove a device is free of faults, and a bad one tells you where to look next.
• Display and audio tests rely on your eyes, your ears, the room and your volume. They are not medical tests of sight or hearing.

How SetupCheck is paid for

The tests are free. The site may show ads from Google AdSense, and we may add affiliate links later. Ads and links never change what a test measures or reports: see the affiliate disclosure, the privacy policy and the terms of use.

Corrections

If a page says something wrong, or a test gives you a result that does not fit what you see, please write to us. We fix mistakes and keep the pages up to date.