Skip to main content

Macro Behavior

Summary​

The macro behavior allows configuring a list of other behaviors to invoke when the macro is pressed and/or released.

Macro Definition​

Each macro you want to use in your keymap gets defined first, then bound in your keymap.

A macro definition looks like:

/ {
    macros {
        zed_em_kay: zed_em_kay {
            compatible = "zmk,behavior-macro";
            #binding-cells = <0>;
            bindings
                = <&macro_press &kp LSHFT>
                , <&macro_tap &kp Z &kp M &kp K>
                , <&macro_release &kp LSHFT>
                ;
        };
    };
};
note

The text before the colon (:) in the declaration of the macro node is the "node label", and is the text used to reference the macro in your keymap.

The macro can then be bound in your keymap by referencing it by the label &zed_em_kay, e.g.:

    raise_layer {
        bindings = <&zed_em_kay>;
    };
note

For use cases involving sending a single keycode with modifiers, for instance ctrl+tab, the key press behavior with modifier functions can be used instead of a macro.

Bindings​

Like hold-taps, macros are created by composing other behaviors, and any of those behaviors can be added to the bindings list, e.g.:

bindings
    = <&to 1>
    , <&bl BL_ON>
    , <&kp Z &kp M &kp K &kp EXCLAMATION>
    ;

Macro Controls​

There are a set of special macro controls that can be included in the bindings list to modify the way the macro is processed.

Binding Activation Mode​

Bindings in a macro are activated differently, depending on the current "activation mode" of the macro.

Available modes:

  • Tap - The default mode; when in this mode, the macro will press, then release, each behavior in the bindings list. This mode is useful for basic keycode output to hosts, i.e. when activating a &kp behavior.
  • Press - In this mode, the macro will only trigger a press on each behavior in the bindings list. This is useful for holding down modifiers for some duration of a macro, e.g. &kp LALT.
  • Release - In this mode, the macro will only trigger a release on each behavior in the bindings list. This is useful for releasing modifiers previously pressed earlier in the macro processing, e.g. &kp LALT.

To modify the activation mode, macro controls can be added at any point in the bindings list.

  • &macro_tap
  • &macro_press
  • &macro_release

A concrete example, used to hold a modifier, tap multiple keys, then release the modifier, would look like:

bindings
    = <&macro_press &kp LSHFT>
    , <&macro_tap &kp Z &kp M &kp K>
    , <&macro_release &kp LSHFT>
    ;

Processing Continuation on Release​

The macro can be paused so that only part of the bindings list is processed when the macro is pressed, and the remainder is processed once the macro itself is released.

To pause the macro until release, use &macro_pause_for_release. For example, this macro will press a modifier and activate a layer when the macro is pressed. Once the macro is released, it will release the modifier and deactivate the layer by releasing the &mo:

bindings
    = <&macro_press &mo 1 &kp LSHFT>
    , <&macro_pause_for_release>
    , <&macro_release &mo 1 &kp LSHFT>
    ;

Wait Time​

The wait time setting controls how long of a delay is introduced between behaviors in the bindings list. The initial wait time for a macro, which is equal to the value of CONFIG_ZMK_MACRO_DEFAULT_WAIT_MS by default, can be set by assigning a value to the wait-ms property of the macro, e.g. wait-ms = <20>;. If you want to update the wait time at any point in the macro bindings list, use &macro_wait_time, e.g. &macro_wait_time 30. A full example:

wait-ms = <10>;
bindings
    = <&kp F &kp A &kp S &kp T>
    , <&macro_wait_time 500>
    , <&kp S &kp L &kp O &kp W>
    ;

Tap Time​

The tap time setting controls how long a tapped behavior is held in the bindings list. The initial tap time for a macro, which is equal to the value of CONFIG_ZMK_MACRO_DEFAULT_TAP_MS by default, can be set by assigning a value to the tap-ms property of the macro, e.g. tap-ms = <20>;. If you want to update the tap time at any point in a macro bindings list, use &macro_tap_time, e.g. &macro_tap_time 30. A full example:

bindings
    = <&macro_tap_time 10>
    , <&kp S &kp H &kp O &kp R &kp T>
    , <&macro_tap_time 500>
    , <&kp L &kp O &kp N &kp G>
    ;

Behavior Queue Limit​

Macros use an internal queue to invoke each behavior in the bindings list when triggered, which has a size of 64 by default. Bindings in "press" and "release" modes correspond to one event in the queue, whereas "tap" mode bindings correspond to two (one for press and one for release). As a result, the effective number of actions processed might be less than 64 and this can cause problems for long macros.

To prevent issues with longer macros, you can change the size of this queue via the CONFIG_ZMK_BEHAVIORS_QUEUE_SIZE setting in your configuration, typically through your .conf file. For example, CONFIG_ZMK_BEHAVIORS_QUEUE_SIZE=512 would allow your macro to type about 256 characters.

Another limit worth noting is that the maximum number of bindings you can pass to a bindings field in the Devicetree is 256, which also constrains how many behaviors can be invoked by a macro.

Parameterized Macros​

Macros can also be "parameterized", allowing them to be bound in your keymap with unique values passed into them, e.g.:

    raise_layer {
        bindings = <&my_one_param_macro A>
    };

Defining Parameterized Macros​

Parameterized macros must be defined using specific values for the compatible and #binding-cells properties, depending on how many parameters they require (up to a maximum of two):

/ {
    macros {
        // 0 params macro
        my_macro: my_macro {
            // ...
            compatible = "zmk,behavior-macro";
            #binding-cells = <0>; // Must be 0
            bindings = /* ... */;
        };

        // 1 param macro
        my_one_param_macro: my_one_param_macro {
            // ...
            compatible = "zmk,behavior-macro-one-param";
            #binding-cells = <1>; // Must be 1
            bindings = /* ... */;
        };

        // 2 params macro
        my_two_param_macro: my_two_param_macro {
            // ...
            compatible = "zmk,behavior-macro-two-param";
            #binding-cells = <2>; // Must be 2
            bindings = /* ... */;
        };
    };
};

Parameters, Bindings and Controls​

There are special macro controls which must be used in order to forward received parameters to the macro's bindings. These controls are "one shot" and will determine how received parameters are used on the very next (non-macro control) behavior in the macro's bindings list.

For example, to pass the first parameter received into a &kp binding, you would use:

bindings = <&macro_param_1to1>, <&kp MACRO_PLACEHOLDER>;

Because kp takes one parameter, you can't simply make the second entry <&kp> in the bindings list. Whatever value you do pass in will be replaced when the macro is triggered, so you can put any value there, e.g. 0, A keycode, etc. To make it very obvious that the parameter there is not actually going to be used, you can use MACRO_PLACEHOLDER which is simply an alias for 0.

The available parameter controls are:

  • &macro_param_1to1 - pass the first parameter of the macro into the first parameter of the next behavior in the bindings list.
  • &macro_param_1to2 - pass the first parameter of the macro into the second parameter of the next behavior in the bindings list.
  • &macro_param_2to1 - pass the second parameter of the macro into the first parameter of the next behavior in the bindings list.
  • &macro_param_2to2 - pass the second parameter of the macro into the second parameter of the next behavior in the bindings list.

Common Patterns​

Below are some examples of how the macro behavior can be used for various useful functionality.

Layer Activation + More​

Macros make it easy to combine a layer behavior, e.g. &mo with another behavior at the same time. Common examples are enabling one or more modifiers when the layer is active, or changing the RBG underglow color.

To achieve this, a combination of a 0ms wait time and splitting the press and release between a &macro_pause_for_release is used:

Layer + modifier​

/**
 * Temporarily switches to a layer (`&mo`) while a modifier is held.
 * Analogous to QMK's `LM()`, using a parameterized macro.
 *
 * Params:
 *  1. Layer to switch to
 *  2. Modifier to press while layer is active
 *
 * Example:
 *  `&lm NUM_LAYER LSHIFT`
 */
lm: lm {
    compatible = "zmk,behavior-macro-two-param";
    wait-ms = <0>;
    tap-ms = <0>;
    #binding-cells = <2>;
    bindings
        = <&macro_param_1to1>
        , <&macro_press &mo MACRO_PLACEHOLDER>
        , <&macro_param_2to1>
        , <&macro_press &kp MACRO_PLACEHOLDER>
        , <&macro_pause_for_release>
        , <&macro_param_2to1>
        , <&macro_release &kp MACRO_PLACEHOLDER>
        , <&macro_param_1to1>
        , <&macro_release &mo MACRO_PLACEHOLDER>
        ;
};

Layer + underglow color​

To trigger a different underglow when the macro is pressed, and when it is released, we use the macro "press" activation mode whenever triggering the &rgb_ug behavior:

wait-ms = <0>;
tap-ms = <0>;
bindings
    = <&macro_press &mo 1>
    , <&macro_tap &rgb_ug RGB_COLOR_HSB(128,100,100)>
    , <&macro_pause_for_release>
    , <&macro_release &mo 1>
    , <&macro_tap &rgb_ug RGB_COLOR_HSB(300,100,50)>;

Keycode Sequences​

The other common use case for macros is to sending sequences of keycodes to the connected host. Here, a wait and tap time of at least 30ms is recommended to avoid having HID notifications grouped at the BLE protocol level and then processed out of order:

wait-ms = <40>;
tap-ms = <40>;
bindings
    = <&kp Z &kp M &kp K>
    , <&kp SPACE>
    , <&kp R &kp O &kp C &kp K &kp S>
    ;

Unicode Sequences​

Many operating systems allow a special sequence to input unicode characters, e.g. Windows alt codes. You can use macros to automate inputting the sequences, e.g. below macro inserts £ on Windows:

wait-ms = <40>;
tap-ms = <40>;
bindings
    = <&macro_press   &kp LALT>
    , <&macro_tap     &kp KP_N0 &kp KP_N1 &kp KP_N6 &kp KP_N3>
    , <&macro_release &kp LALT>
    ;

Convenience C Macro​

To avoid repetition or possible typos when declaring a zero parameter macro, a convenience C macro, named ZMK_MACRO(name, props) can be used to simplify things:

    ZMK_MACRO(my_zero_param_macro,
        wait-ms = <30>;
        tap-ms = <40>;
        bindings = <&kp Z &kp M &kp K>;
    )
note

ZMK_MACRO() only supports declaring non-parameterized (zero parameter) macros; parameterized declarations are not currently supported.

This can be used instead of a complete macro definition. During the firmware build process, the example above would produce the complete macro definition below:

    my_zero_param_macro: my_zero_param_macro {
        compatible = "zmk,behavior-macro";
        #binding-cells = <0>;
        wait-ms = <30>;
        tap-ms = <40>;
        bindings = <&kp Z &kp M &kp K>;
    };

Using the C macro is entirely optional, and is provided only as a convenience.