TypeCrt vs Monkeytype: an honest comparison

We built one of these two. That is a conflict of interest, so every claim below comes with a way for you to check it without taking our word for anything.

Comparison Updated 12 August 2026· 5 min· By the TypeCrt developer

Correction, August 2026. An earlier version of this article claimed Monkeytype “shows ads to non-logged-in users” and listed zero-ads as a TypeCrt advantage over it. That was wrong. Monkeytype's advertisements are optional and can be switched off in settings without paying — the project's own README describes it as “minimalistic design, with optional advertisements.” The row has been corrected below. We would rather lose a talking point than keep a false one.

Contents
  1. The short answer
  2. Feature by feature
  3. Where Monkeytype is genuinely better
  4. Where TypeCrt is genuinely better
  5. Verify any of this yourself
  6. Which one you should use

The short answer

Monkeytype is the better measurement instrument. TypeCrt is the better training instrument. That is the whole comparison, and everything below is detail.

Monkeytype tells you how fast you type, in more languages, against a global leaderboard, with source code you can read. TypeCrt tells you which keys are costing you the speed and then generates practice weighted toward those keys. If you already know what to fix, you do not need the second thing. Most people do not know what to fix.

A typing test measures. A practice tool intervenes. Most tools claiming to be both are really only the first.

Feature by feature

Verified 12 August 2026 against Monkeytype's public repository and live site. Where we could not verify a number, we say so instead of estimating.

FeatureTypeCrtMonkeytype
PriceFreeFree
AdvertisingNone at allOptional, off by default in settings
Account requiredNoNo
Optional account syncYesYes
Open sourceNo — proprietaryYes — GPL-3.0
Global leaderboardsNoneYes, English time tests
LanguagesEnglish onlyMany
Themes80, each listed at /docs/themesLarger, community-contributed
Adaptive weak-key practiceYes — KeyForgeNo
Per-key keyboard heatmapYesNo
CRT scanlines and phosphor glowYesNo
Documented settings35, listed at /docs/settingsExtensive
Time modes15 / 30 / 60 / 120 s15 / 30 / 60 / 120 s
Word modes10 / 25 / 50 / 10010 / 25 / 50 / 100
Quote, Zen and custom textYesYes
Consistency metricYes — formula publishedYes
Speed-over-time analysisYes — WPM per hour of practiceNo
Test modifiers / funboxNoYes

Where Monkeytype is genuinely better

Four of these are structural. We are not going to close them soon, and pretending otherwise would make the rest of this page worthless.

Where TypeCrt is genuinely better

KeyForge — adaptive practice

Monkeytype measures. It does not diagnose. KeyForge scores every key from your own typing history using response time and error rate, keeps a rolling average so one bad test does not distort it, and then generates practice text weighted toward the keys scoring worst. Weak keys appear more often; solid keys appear less.

The mechanism, including the exact confidence formula and the order letters unlock in, is written up at /docs/keyforge rather than left as a marketing word.

The per-key heatmap

After every test, the keyboard is redrawn with each key coloured by your speed and accuracy on it. It takes about two seconds to see that, say, your left ring finger is the problem — which is the sort of thing you can practise for a year without noticing from a single WPM number.

CRT rendering

Scanlines, phosphor bloom and screen curvature, rendered live rather than as a static filter. It is aesthetic rather than functional and we make no claim that it makes you faster. It is simply not available anywhere else, and for some people it is the reason to open a typing test at all.

No advertising, on any tier

Not a dig at Monkeytype, whose ads are optional too. It is a statement about what TypeCrt will do later: there is no ad slot in the codebase and no plan to add one.

Verify any of this yourself

Comparisons written by a vendor are worth exactly as much as their falsifiability. Here is how to falsify each of the four claims above that actually matter.

Check the adaptive claim
  1. Open typecrt.com and run three or four normal tests, typing naturally — mistakes included.
  2. Switch to KeyForge. The generated text will over-represent the keys you were slowest and least accurate on.
  3. Compare against the per-key panel. If a key you are demonstrably bad at is not appearing more often, the claim is false and we want the bug report.
Check the no-ads claim
  1. Open TypeCrt, press F12, and open the Network tab.
  2. Reload and read the request list. No ad network, no ad exchange, no tracking pixel appears — the full list of third parties the site contacts is published on the privacy page.
Check the theme count

The themes page is generated directly from the theme definitions in the source at build time, not written by hand. Count the rows. If it says 80, there are 80.

Check the metric definitions

Both sites use the standard five-characters-equals-one-word convention, so scores are broadly comparable. Ours is written out with worked examples at /docs/metrics; theirs is readable in their source. Run a 60-second test on each within a few minutes and the two numbers should land close together. If they do not, one of us has a bug, and we would like to know if it is us.

Which one you should use

Use Monkeytype if you type in a language other than English, want to read or self-host the source, care about global rankings, or want the largest theme library and the biggest community.

Use TypeCrt if you have plateaued and do not know why, want practice that targets your specific weak keys, want a per-key heatmap after each test, or want the CRT aesthetic.

Use both. They are free, neither requires an account, and they are not really competing for the same twenty minutes. Practise weak keys here; benchmark yourself there. That is genuinely what we would advise a friend, and it is what makes the rest of this page believable.

See where your weak keys are

One 60-second test produces the per-key heatmap. No account, no ads, nothing to install.

Start a typing test →
TypeCrt Blog Docs About Privacy