#typing #tui #wpm #comfy #word #script #color

bin+lib smokey

Comfy terminal based typing test

10 releases

0.3.4 Mar 15, 2022
0.3.3 Mar 10, 2022
0.3.2 Dec 4, 2021
0.3.0 Nov 21, 2021
0.1.0 May 26, 2021

#400 in Hardware support

MIT license



Comfy terminal based typing test


With cargo

cargo install smokey

Building from source

git clone https://github.com/ukmrs/smokey && cd smokey
cargo build --release

then copy target/release/smokey to a known location


Typing Test Screen

TABReset the current test
ESCOpen the settings
CTRL + CExit
CTRL + BackspaceDelete a word

Settings Screen

TABStart a new test
h j k l / Arrow KeysMovement
d / ESCDeselect
s / ENTERSelect
q / ESC / CTRL + CExit

Results Screen

TABStart a new test
sOpen the settings
q / ESC / CTRL + CExit

word lists

Smokey ships with a sizeable english word list (~60_000 words) which on linux can be found in


Otherwise location can be found with the --storage flag.

More lists can be added to the folder. Smokey expects a list sorted by word frequency with each word separated by a newline character. Other languages are not provided but most of the time can be easily DIYed.

Suggestions/Examples for word sources


Grab Lexique382.zip and unzip it to find Lexique382.tsv. If you have xsv installed, here is an almost one liner to convert it to smokey-friendly format:

xsv sort -s freqlivres -N -R Lexique382.tsv | xsv select ortho > french
sed '1d' french > tmpfile && mv tmpfile french

The only purpose of the sed command is to delete the first line which will be "ortho" - the column name. That could be done manually but I included it for convenience.


Grab Otwarty słownik frekwencyjny leksemów pdftotext it, sort it and clean it by a short script. Godspeed mój przyjacielu.

Supported languages/scripts

Smokey should handle all simple scripts like

  • latin derivatives
  • cyrylic
  • greek
  • etc

Complex scripts that require mulitple inputs for one glyph like Hangul won't work. The same goes for right_to_left scripts.

English word list

The list contains around 60 000 words. It is derived from 1/3 million most frequent English words compiled by Peter Norvig. I filtred it using python bindings of Enchant and checked against the MauriceButler/badwords and LDNOOBW. I kept "sex" though. Otherwise it wouldn't be fair to plants who just cross-pollinate without causing too much of a ruckus. Future me here, I forgot about accursed pollen allergies, I might reconsider my stance on this.


There is a scripts directory in which you can throw... scripts. Just chmod +x, add a shebang, slap it in there and the output will be converted to a typing test. There is an example python script shipped with smokey that produces gibberish but of course you can add whatever else. For instance, I use a script that fetches me a random quote from a local database. You can snatch some from Monkeytype or TypeRacer to name a few.


You can create smokey.toml configuration file that allows to change colors or set default test settings. On linux:


Other OS:

smokey --config

Example smokey.toml

For colors you can use either hex codes or standard colors (supported names).

# test colors
todo = "grey"
done = "#96BB7C"
mistake = "#C64756"

# settings colors
active = "#93a1bf"
hover = "#aa78bf"

# default test settings
name = "english"
mods = ["punctuation", "numbers"]
len = 20
pool = 60000

Run history

Runs are saved to a sqlite database, on linux you can find it here:


For now it is only used to fetch and compare against your record wpm. In the very near future there will be a nice way to explore the history, but I haven't implemented that yet, soz.


~573K SLoC