One very basic operation is opening a new terminal. By default, the keybinding for this is Mod1+Enter, that is Alt+Enter in the default configuration. By pressing Mod1+Enter, a new terminal will be opened. It will fill the whole space available on your screen.

It is important to keep in mind that i3 uses a table to manage your windows. At the moment, you have exactly one column and one row which leaves you with one cell. In this cell there is a container, which is where your new terminal is opened.

If you now open another terminal, you still have only one cell. However, the container in that cell holds both of your terminals. So, a container is just a group of clients with a specific layout. Containers can be resized by adjusting the size of the cell that holds them.

To move the focus between the two terminals, you use the direction keys which you may know from the editor vi. However, in i3, your homerow is used for these keys (in vi, the keys are shifted to the left by one for compatibility with most keyboard layouts). Therefore, Mod1+J is left, Mod1+K is down, Mod1+L is up and Mod1+; is right. So, to switch between the terminals, use Mod1+K or Mod1+L.

To create a new row/column (and a new cell), you can simply move a terminal (or any other window) in the direction you want to expand your table. So, let’s expand the table to the right by pressing Mod1+Shift+;.

A container can have the following modes:

To switch modes, press Mod1+e for default, Mod1+h for stacking and Mod1+w for tabbed.

To display a window fullscreen or to go out of fullscreen mode again, press Mod1+f.

There is also a global fullscreen mode in i3 in which the client will use all available outputs. To use it, or to get out of it again, press Mod1+Shift+f.

Aside from opening applications from a terminal, you can also use the handy dmenu which is opened by pressing Mod1+v by default. Just type the name (or a part of it) of the application which you want to open. The application typed has to be in your $PATH for this to work.

Additionally, if you have applications you open very frequently, you can create a keybinding for starting the application directly. See the section "Configuring i3" for details.

If an application does not provide a mechanism for closing (most applications provide a menu, the escape key or a shortcut like Control+W to close), you can press Mod1+Shift+q to kill a window. For applications which support the WM_DELETE protocol, this will correctly close the application (saving any modifications or doing other cleanup). If the application doesn’t support the WM_DELETE protocol your X server will kill the window and the behaviour depends on the application.

Workspaces are an easy way to group a set of windows. By default, you are on the first workspace, as the bar on the bottom left indicates. To switch to another workspace, press Mod1+num where num is the number of the workspace you want to use. If the workspace does not exist yet, it will be created.

A common paradigm is to put the web browser on one workspace, communication applications (mutt, irssi, …) on another one, and the ones with which you work, on the third one. Of course, there is no need to follow this approach.

If you have multiple screens, a workspace will be created on each screen at startup. If you open a new workspace, it will be bound to the screen you created it on. When you switch to a workspace on another screen, i3 will set focus to that screen.

To move a window to another workspace, simply press Mod1+Shift+num where num is (like when switching workspaces) the number of the target workspace. Similarly to switching workspaces, the target workspace will be created if it does not yet exist.

To resize columns or rows, just grab the border between the two columns/rows and move it to the wanted size. Please keep in mind that each cell of the table holds a container and thus you cannot horizontally resize single windows. If you need applications with different horizontal sizes, place them in seperate cells one above the other.

See [resizingconfig] for how to configure i3 to be able to resize columns/rows with your keyboard.

To restart i3 inplace (and thus get into a clean state if there is a bug, or to upgrade to a newer version of i3) you can use Mod1+Shift+r. Be aware, though, that this kills your current layout and all the windows you have opened will be put in a default container in only one cell. Saving layouts will be implemented in a later version.

To cleanly exit i3 without killing your X server, you can use Mod1+Shift+e.

Snapping is a mechanism to increase/decrease the colspan/rowspan of a container. Colspan/rowspan is the number of columns/rows a specific cell of the table consumes. This is easier explained by giving an example, so take the following layout:

To use the full size of your screen, you can now snap container 3 downwards by pressing Mod1+Control+k (or snap container 2 rightwards).

Floating mode is the opposite of tiling mode. The position and size of a window are not managed by i3, but by you. Using this mode violates the tiling paradigm but can be useful for some corner cases like "Save as" dialog windows, or toolbar windows (GIMP or similar).

You can enable floating mode for a window by pressing Mod1+Shift+Space. By dragging the window’s titlebar with your mouse you can move the window around. By grabbing the borders and moving them you can resize the window. You can also do that by using the [floating_modifier].

For resizing floating windows with your keyboard, see [resizingconfig].

Floating windows are always on top of tiling windows.

It is possible and recommended to use comments in your configuration file to properly document your setup for later reference. Comments are started with a # and can only be used at the beginning of a line:

Examples:

i3 uses X core fonts (not Xft) for rendering window titles and the internal workspace bar. You can use xfontsel(1) to generate such a font description. To see special characters (Unicode), you need to use a font which supports the ISO-10646 encoding.

Syntax:

Examples:

A keyboard binding makes i3 execute a command (see below) upon pressing a specific key. i3 allows you to bind either on keycodes or on keysyms (you can also mix your bindings, though i3 will not protect you from overlapping ones).

My recommendation is: If you often switch keyboard layouts but you want to keep your bindings in the same physical location on the keyboard, use keycodes. If you don’t switch layouts, and want a clean and simple config file, use keysyms.

Syntax:

Examples:

Available Modifiers:

To move floating windows with your mouse, you can either grab their titlebar or configure the so called floating modifier which you can then press and click anywhere in the window itself to move it. The most common setup is to use the same key you use for managing windows (Mod1 for example). Then you can press Mod1, click into a window using your left mouse button, and drag it to the position you want.

When holding the floating modifier, you can resize a floating window by pressing the right mouse button on it and moving around while holding it. If you hold the shift button as well, the resize will be proportional.

Syntax:

Examples:

This option determines in which mode new containers will start. See also [stack-limit].

Syntax:

Examples:

This option determines which border style new windows will have.

Syntax:

Examples:

As you learned in the section about keyboard bindings, you will have to configure lots of bindings containing modifier keys. If you want to save yourself some typing and be able to change the modifier you use later, variables can be handy.

Syntax:

Examples:

Variables are directly replaced in the file when parsing. There is no fancy handling and there are absolutely no plans to change this. If you need a more dynamic configuration you should create a little script which generates a configuration file and run it before starting i3 (for example in your .xsession file).

It is recommended that you match on window classes wherever possible because some applications first create their window, and then worry about setting the correct title. Firefox with Vimperator comes to mind. The window starts up being named Firefox, and only when Vimperator is loaded does the title change. As i3 will get the title as soon as the application maps the window (mapping means actually displaying it on the screen), you’d need to have to match on Firefox in this case.

You can prefix or suffix workspaces with a ~ to specify that matching clients should be put into floating mode. If you specify only a ~, the client will not be put onto any workspace, but will be set floating on the current one.

Syntax:

Examples:

Note that the arrow is not required, it just looks good :-). If you decide to use it, it has to be a UTF-8 encoded arrow, not "→" or something like that.

By using the exec keyword outside a keybinding, you can configure which commands will be performed by i3 on initial startup (not when restarting i3 in-place however). These commands will be run in order.

Syntax:

Examples:

If you assign clients to workspaces, it might be handy to put the workspaces on specific screens. Also, the assignment of workspaces to screens will determine which workspace i3 uses for a new screen when adding screens or when starting (e.g., by default it will use 1 for the first screen, 2 for the second screen and so on).

Syntax:

The output is the name of the RandR output you attach your screen to. On a laptop, you might have VGA1 and LVDS1 as output names. You can see the available outputs by running xrandr --current.

Examples:

If you always have a certain arrangement of workspaces, you might want to give them names (of course UTF-8 is supported):

Syntax:

For more details about the output part of this command, see above.

Examples:

You can change all colors which i3 uses to draw the window decorations and the bottom bar.

Syntax:

Where colorclass can be one of:

You can also specify the color to be used to paint the background of the client windows. This color will be used to paint the window on top of which the client will be rendered.

Syntax:

Only clients that do not cover the whole area of this window expose the color used to paint it. If you use a color other than black for your terminals, you most likely want to set the client background color to the same color as your terminal program’s background color to avoid black gaps between the rendered area of the termianal and the i3 border.

Colors are in HTML hex format (#rrggbb), see the following example:

Examples:

Note that for the window decorations, the color around the child window is the background color, and the border color is only the two thin lines at the top of the window.

i3 uses unix sockets to provide an IPC interface. This allows third-party programs to get information from i3, such as the current workspaces (to display a workspace bar), and to control i3.

To enable it, you have to configure a path where the unix socket will be stored. The default path is ~/.i3/ipc.sock.

Examples:

You can then use the i3-msg application to perform any command listed in the next section.

If you have a setup where your mouse usually is in your way (like a touchpad on your laptop which you do not want to disable completely), you might want to disable focus follows mouse and control focus only by using your keyboard. The mouse will still be useful inside the currently active window (for example to click on links in your browser window).

Syntax:

Examples:

The internal workspace bar (the thing at the bottom of your screen) is very simple — it does not provide a way to display custom text and it does not offer advanced customization features. This is intended because we do not want to duplicate functionality of tools like dzen2, xmobar and so on (they render bars, we manage windows). Instead, there is an option which will turn off the internal bar completely, so that you can use a separate program to display it (see i3-wsbar, a sample implementation of such a program):

Syntax:

Examples:

To change the layout of the current container to stacking, use s, for default use d and for tabbed, use T. To make the current client (!) fullscreen, use f, to make it span all outputs, use fg, to make it floating (or tiling again) use t:

Examples:

To change the focus, use one of the h, j, k and l commands, meaning left, down, up, right (respectively). To focus a container, prefix it with wc. To focus a screen, prefix it with ws.

The same principle applies for moving and snapping: just prefix the command with m when moving and with s when snapping:

Examples:

To change to a specific workspace, the command is just the number of the workspace, e.g. 1 or 3. To move the current client to a specific workspace, prefix the number with an m.

You can also switch to the next and previous workspace with the commands nw and pw, which is handy, for example, if you have workspace 1, 3, 4 and 9 and you want to cycle through them with a single key combination.

Examples:

If you want to resize columns/rows using your keyboard, you can use the resize command, I recommend using it inside a so called mode:

Often when in a multi-monitor environment, you want to quickly jump to a specific window. For example, while working on workspace 3 you may want to jump to your mail client to email your boss that you’ve achieved some important goal. Instead of figuring out how to navigate to your mailclient, it would be more convenient to have a shortcut.

Syntax:

You can either use the same matching algorithm as in the assign command (see above) or you can specify the position of the client if you always use the same layout.

Examples:

This feature is like the jump feature: It allows you to directly jump to a specific window (this means switching to the appropriate workspace and setting focus to the windows). However, you can directly mark a specific window with an arbitrary label and use it afterwards. You do not need to ensure that your windows have unique classes or titles, and you do not need to change your configuration file.

As the command needs to include the label with which you want to mark the window, you cannot simply bind it to a key. i3-input is a tool created for this purpose: It lets you input a command and sends the command to i3. It can also prefix this command and display a custom prompt for the input dialog.

Syntax:

Examples:

Alternatively, if you do not want to mess with i3-input, you could create seperate bindings for a specific set of labels and then only use those labels.

This mechanism can be thought of as the opposite of the jump command. It travels the focus stack and jumps to the window which had focus previously.

Syntax:

Where number by default is 1 meaning that the next client in the focus stack will be selected.

The special values have the following meaning:

To change the border of the current client, you can use bn to use the normal border (including window title), bp to use a 1-pixel border (no window title) and bb to make the client borderless. There is also bt which will toggle the different border styles.

Examples:

If you have a single container with a lot of windows inside it (say, more than 10), the default layout of a stacking container can get a little unhandy. Depending on your screen’s size, you might end up seeing only half of the titlebars for each window in the container.

Using the stack-limit command, you can limit the number of rows or columns in a stacking container. i3 will create columns or rows (depending on what you limited) automatically as needed.

Syntax:

Examples:

You can make i3 reload its configuration file with reload. You can also restart i3 inplace with the restart command to get it out of some weird state (if that should ever happen) or to perform an upgrade without having to restart your X session. However, your layout is not preserved at the moment, meaning that all open windows will end up in a single container in default layout after the restart. To exit i3 properly, you can use the exit command, however you don’t need to (simply killing your X session is fine as well).

Examples:

To help you get going if you have never used multiple monitors before, here is a short overview of the xrandr options which will probably be of interest to you. It is always useful to get an overview of the current screen configuration. Just run "xrandr" and you will get an output like the following:

Several things are important here: You can see that LVDS1 is connected (of course, it is the internal flat panel) but VGA1 is not. If you have a monitor connected to one of the ports but xrandr still says "disconnected", you should check your cable, monitor or graphics driver.

The maximum resolution you can see at the end of the first line is the maximum combined resolution of your monitors. By default, it is usually too low and has to be increased by editing /etc/X11/xorg.conf.

So, say you connected VGA1 and want to use it as an additional screen:

This command makes xrandr try to find the native resolution of the device connected to VGA1 and configures it to the left of your internal flat panel. When running "xrandr" again, the output looks like this:

Please note that i3 uses exactly the same API as xrandr does, so it will see only what you can see in xrandr.

See also [presentations] for more examples of multi-monitor setups.

There are several things to configure in i3 which may be interesting if you have more than one monitor:

A very common thing amongst users of exotic window managers is a status line at some corner of the screen. It is an often superior replacement to the widget approach you have in the task bar of a traditional desktop environment.

If you don’t already have your favorite way of generating such a status line (self-written scripts, conky, …), then i3status is the recommended tool for this task. It was written in C with the goal of using as few syscalls as possible to reduce the time your CPU is woken up from sleep states.

Regardless of which application you use to generate the status line, you want to make sure that the application does one of the following things:

The planned solution for this problem is to make the workspace bar optional and switch to a third party application completely (dzen2 for example) which will then contain the workspace bar.

When giving a presentation, you typically want the audience to see what you see on your screen and then go through a series of slides (if the presentation is simple). For more complex presentations, you might want to have some notes which only you can see on your screen, while the audience can only see the slides.