Memory Test Online is a free set of memory tests you can take in your browser, built to be genuinely useful rather than a wall of ads. This page explains who runs the site, how the tests are built, and where every number comes from. The short version: no score on this site is invented, and each test traces to published research.
Who’s behind the site
The site is built and run by Tim, an independent web developer who also runs timmyships.com. This is a solo project, not a company or a medical service. That matters for how you should read it: the tests are a well-made tool for curiosity and self-tracking, put together carefully, but the site doesn’t pretend to be a clinic.
Where the site adds real value is in doing the homework most free memory-test pages skip. The science is sourced from published memory research, the scores are computed rather than guessed, and the honest limits of each test are stated plainly.
How the tests work
Five test engines power all the tests on the site:
- A number engine, where a growing digit string flashes and you type it back. It powers the number memory, digit span, and short-term memory tests.
- A visual engine, where squares light on a grid and you tap them back. It powers the visual, photographic, and eidetic memory tests.
- A sequence engine, where tiles light in order and you repeat them.
- A verbal engine, where you mark words as seen or new.
- A chimp engine, where numbered tiles hide and you click them in order.
Framing several tests on one engine is deliberate. “Digit span” and “number memory” measure the same underlying skill under different names people actually search for, so the same reliable task serves both.
Where the scores come from
This is the part we take seriously. Every average, percentile, and band shown on the site is calculated by the same scoring engine that runs the live tests, checked against published norms where they exist. Nothing is typed in by hand.
That design makes a certain kind of error impossible. A page can’t claim a “typical score” that disagrees with what the tool actually produces, because both read from the same source. When published academic norms exist for a task, like digit span, the site uses them and says so. Where a figure is a modeled or typical range rather than a hard published number, it’s labeled that way. The goal is that you can always tell where a number came from.
An honest note on what these tests are
These are self-tests for interest and practice. They are not medical or diagnostic tools, and they can’t screen for any condition. A single low score usually just means you were tired or distracted. If you’re worried about your memory, see a doctor rather than an online test. You can read more of that guidance on the memory test for seniors page.
Contact
Questions, corrections, or a norm you think we’ve got wrong? Email info@memorytestonline.com. Corrections to the science are genuinely welcome; getting the numbers right is the whole point of the site.
Frequently asked questions
Is Memory Test Online free?
Yes. Every test is free, needs no account, and works on phones and computers. Your scores save on your own device so you can track progress. The site is ad-free at launch, built as a clean tool rather than a page stuffed with ads.
Are the memory tests scientifically accurate?
Each test accurately measures the specific skill it targets, using tasks drawn from established memory research. They’re built for interest and self-tracking, not clinical diagnosis. Where published norms exist, the site uses them; where a figure is a modeled range, it says so. They can’t diagnose any medical condition.
Who created Memory Test Online?
Tim, an independent web developer who also runs timmyships.com, built and runs the site as a solo project. It isn’t a company or a medical service. The value comes from careful sourcing: computed scores, cited research, and honest statements about what each test can and can’t tell you.
How do you make sure the scores are accurate?
Every average, percentile, and band is computed by the same engine that runs the live tests, checked against published norms where they exist. No number is hand-written, so a page can’t drift from what the tool produces. Modeled ranges are labeled as such, so you can always tell how solid a figure is.