A screencast tool to display your keys, inspired by Screenflick.

This is an almost-complete rewrite of screenkey 0.2, featuring:

• Several keyboard translation methods
• Key composition/input method support
• Configurable font/size/position
• Highlighting of recent keystrokes
• Improved backspace processing
• Normal/Emacs/Mac caps modes
• Multi-monitor support
• Dynamic recording control
• Switch for visible shift and modifier sequences only
• Repeats compression
• Countless bug fixes

Execute without installation:

./screenkey

To install:

sudo ./setup.py install

Dependencies:

• Python 2.7 (Python 3 support yet incomplete)
• PyGTK
• setuptools (build only)
• DistUtils-Extra (build only)
• slop (https://github.com/naelstrof/slop)
• FontAwesome (for multimedia symbols)

Install dependencies (on Debian/Ubuntu):

sudo apt-get install python-gtk2 python-setuptools python-setuptools-git python-distutils-extra

You can also install "screenkey" via ArchLinux's AUR package:

https://aur.archlinux.org/packages/screenkey

Display time:

Persistence (in seconds) of the output window after typing has stopped. Defaults to 2.5 seconds. When the window is persistent, display time still controls the time before the text is cleared.

Persistent window:

Forces the output window to be always visible, irregardless of typing activity. Mostly useful for interactive window placement and/or "fixed" positioning.

Screen:

Physical screen/monitor used for the output window.

Position:

Position of the output window. The position is normally relative to the chosen screen. If a window has been selected with "Select window/region", the position becomes relative to the window. If "fixed" is chosen, the output window's position and size are specified explicitly. See Interactive placement for more details.

Font:

Font used for the output window. A scalable font and wide Unicode coverage is required (the DejaVu family is highly recommended).

Size:

Size of the font used in the output window. Chooses proportionally between 8/12/24% of the screen size. When "fixed" positioning is used, size is ignored and the font will fill the available height of the output window.

Keyboard mode:

Choose the translation method of keyboard events.

"Composed" attempts to show only the final results of key composition. Dead keys and any intermediate output during composition is not shown. Currently works correctly with XIM/IBUS, but only for on-the-spot editing. It can cause problems with complex input methods (support for wider compatibility is underway).

"Translated" shows the result of each keypress on the keyboard, accounting for the current keyboard locale and modifiers, but not composition. Pressing a dead key followed by a letter will show both keys.

"Raw" shows which key caps were pressed on the keyboard, without translation. For example, typing "!" (which is often located on top of the key "1") requires pressing "Shift+1", which is what this output mode shows. "Backspace mode", "Always visible Shift" and "Modifiers only" have no effect in this mode.

"Keysyms" shows the keysyms ("symbolic" names) of each pressed key as received by the server. Mostly useful for debugging.

Backspace mode:

Controls the effect of "backspace" on the text in the output window.

"Normal" always inserts a backspace symbol in the output window.

"Baked" simulates the effect of backspace in the text only if the last keypress is a regular letter and no caret movement has been detected. In any other case, a backspace symbol is inserted instead.

"Full" is similar to "baked", but will eat through several other, less safe keys, such as tabs and returns.

Modifiers mode:

Select how modifiers keys (such as Control, Alt) are displayed in the output window. "Normal" uses traditional PC names (Ctrl+A) while "Mac" uses Mac symbols directly (⌘+A). The "Emacs" mode will display Emacs-style shortened keyboard sequences (C-A).

Show Modifier sequences only:

Only show modifier/control sequences in the output window. Bare, shifted or translated letters are not shown.

Always show Shift:

Shift is normally hidden when the control sequence includes a letter that can differentiate between a shifted/non-shifted key. For example, Shift + "Control+a" is normally shown just as "Control+A" (notice the capital "A").

When "Always show Shift" is used, Shift is always included in modifier sequences, if pressed. Has no effect when using the "Emacs" modifiers mode.

Show Whitespace characters:

Convert regular whitespace characters (tabs and spaces) to a visible representation instead of showing a blank. Newlines are also hidden when unambiguous in multiline mode.

Compress repeats:

When enabled, contiguous repeated sequences are truncated after the requested threshold. A counter of total occurrences is shown instead, which is generally more legible.

To disable screenkey while recording (for example, during password prompts), press both control keys, or both shift keys, or both alt keys at the same time.

Press the same combination again to resume it.

This has the same effect of toggling the state from the system tray icon, but it's completely stealth: there's no feedback that screenkey is being switched on/off.

If you need the viewer to focus on a sentence you just typed, you can press a silent modifier (such as Shift, or Control) to keep the output window visible a little longer without prolonging the default timeout.

screenkey is normally positioned on the top/center/bottom part of the screen.

If you're recording a screencast only for a specific application, you can click on "Select window/region" to select on which window the output should be overlaid (slop must be installed for this task). When a window has been selected, top/center/bottom refer to the window's contents. Press "Reset" to restore the original behavior.

When "fixed" is chosen, the position of the output is specified directly. The cursor turns immediately into a crossbar: drag over the desired screen region (where the text should appear), or press "Esc" to abort. Again, press "Reset" to restore the original behavior.

The "geometry" argument follows the standard X11 geometry format (WxH[+X+Y]) and can be provided by slop, which allows to select windows and/or drag over the desired region interactively without the need of calculating the coordinates manually.

When a geometry argument has been provided, the position (top/middle/bottom) becomes relative to the selected rectangle. For example, to overlay screenkey on top of an existing window, you can simply do:

./screenkey -g $(slop -n -f '%g')

To set the actual text rectangle instead, use "fixed" positioning. Using slop, you can combine both and simply drag the desired rectangle during selection:

./screenkey -p fixed -g $(slop -n -f '%g')

The default font is "Sans Bold", which is usually mapped to "DejaVu Sans" on most Linux installations (look for the ttf-dejavu package). It's a good all-around font which provides all the required glyphs and has excellent readability.

For screencasts about programming, we recommend "DejaVu Sans Mono Bold" instead, which provides better differentiation among similar letterforms (0/O, I/l, etc).

"screenkey" supports several multimedia keys. To display them with symbols instead of text abbreviations, FontAwesome needs to be installed.

On Debian/Ubuntu, the font is available in the fonts-font-awesome package. On Arch Linux the package is instead ttf-font-awesome.

"screenkey" should work correctly by default with any tiling window manager.

The original version of screenkey used to require customization for the output window to work/float correctly. These settings are no longer required with this fork, and can be safely removed.

If you don't have a system tray, you can either configure screenkey through command line flags or use --show-settings to test the configuration interactively.

To get transparency you need a compositor to be running. For example, "compton" or "unagi" are popular for their low impact on performance, but "xcompmgr" also works correctly without any additional configuration.

If you're recording a screencast where almost all editing is already visible (for example, in vi or most other text editors), consider using a bigger screen font instead, so that the viewer can read the text directly while the program is being used.

If the control sequences you're typing are rare, you might even want to spell what you're doing instead of obscuring the screen with the typing output.

When doing screencasts involving a lot of mouse activity, or which require holding down modifiers to perform other mouse actions, key-mon might be a good companion to screenkey, or replace it entirely.

key-mon can be configured to show the state of key modifiers continuously and circle the location of mouse clicks ("visible click"). key-mon and screenkey complete each-other and can be used at the same time.

"screenkey" can be found at https://www.thregr.org/~wavexx/software/screenkey/

screenkey's GIT repository is publicly accessible at:

https://github.com/wavexx/screenkey

• Benjamin Chrétien
• Dmitry Bushev
• Doug Patti
• Igor Bronovskyi
• Ivan Makfinsky
• Jacob Gardner
• Muneeb Shaikh
• Stanislav Seletskiy
• farrer (launchpad)
• zhum (launchpad)
• 伊冲