Tap-dance keys are general purpose, multi-use keys, which trigger a different action based on the number of times they were tapped in sequence. As an example to make this clearer, one can have a key that inputs A when tapped once, inputs B when tapped twice, and lights up the keyboard in Christmas colors when tapped a third time.

This behaviour is most useful in cases where we have a number of things we perform rarely, where tapping a single key repeatedly is not counter-productive. Such cases include - for example - multimedia forward / backward keys: forward on single tap, backward on double. Of course, one could use modifiers to achieve a similar effect, but that’s two keys to use, this is only one. We can also hide some destructive functionality behind a number of taps: reset the keyboard after 4 taps, and light up LEDs in increasingly frightful colors until then.

How does it work?

To not interfere with normal typing, tap-dance keys have two ways to decide when to call an action: they either get interrupted, or they time out. Every time a tap-dance key is pressed, the timer resets, so one does not have to finish the whole tapping sequence within a short time limit. The tap-dance counter continues incrementing until one of these cases happen.

When a tap-dance key is pressed and released, and nothing is pressed on the keyboard until the timeout is reached, then the key will time out, and trigger an action. Which action, depends on the number of times it has been tapped up until this point.

When a tap-dance key is pressed and released, and another key is hit before the timer expires, then the tap-dance key will trigger an action first, perform it, and only then will the firmware continue handling the interrupting key press. This is to preserve the order of keys pressed.

In both of these cases, the user-defined tapDanceAction() function will be called, with tap_dance_index set to the index of the tap-dance action (as set in the keymap), the tap_count, and tap_dance_action set to one of the following values:

  • kaleidoscope::plugin::TapDance::Hold, if the tap-dance key is still being held when its timeout expires.

  • kaleidoscope::plugin::TapDance::Timeout, if the tap-dance key has been released when its timeout expires.

  • kaleidoscope::plugin::TapDance::Interrupt, if another key is pressed before the tap-dance key’s timeout expires.

These actions allow us to create sophisticated tap-dance setups, where one can tap a key twice and hold it, and have it repeat, for example.

There is one additional value the tap_dance_action parameter can take: kaleidoscope::plugin::TapDance::Tap. It is called with this argument for each and every tap, even if no action is to be triggered yet. This is so that we can have a way to do some side-effects, like light up LEDs to show progress, and so on.

Using the plugin

To use the plugin, we need to include the header, and declare the behaviour used. Then, we need to place tap-dance keys on the keymap. And finally, we need to implement the tapDanceAction function that gets called each time an action is to be performed.

#include <Kaleidoscope.h>
#include <Kaleidoscope-TapDance.h>

// Somewhere in the keymap:

// later in the Sketch:
void tapDanceAction(uint8_t tap_dance_index, KeyAddr key_addr, uint8_t tap_count,
                    kaleidoscope::plugin::TapDance::ActionType tap_dance_action) {
  switch (tap_dance_index) {
  case 0:
    return tapDanceActionKeys(tap_count, tap_dance_action,
                              Consumer_ScanNextTrack, Consumer_ScanPreviousTrack);


void setup() {
  Kaleidoscope.setup ();

Keymap markup


A key that acts as a tap-dance key. The actions performed depend on the implementation for the id index within the [tapDanceActions][tdactions] function.

The id parameter here is what will be used as tap_dance_index in the handler function.

Plugin methods

The plugin provides a TapDance object, but to implement the actions, we need to define a function (tapDanceAction) outside of the object. A handler, of sorts. Nevertheless, the plugin provides one macro that is particularly useful: tapDanceActionKeys. Apart from that, it provides only one configuration method:


Set the number of milliseconds to wait before a tap-dance sequence times out. Once the sequence timed out, the action for it will trigger, even without an interruptor. Defaults to 5, and the timer resets with every tap of the same

tapDanceActionKeys(tap_count, tap_dance_action, keys...)

Sets up an action where for each subsequent tap, a different key will be chosen from the list of keys supplied in the keys... argument.

If we have Key_A and Key_B in the list, then, if tapped once, this function will input A, but when tapped twice, will input B.

When all our actions are just different keys, this is a very handy macro to use.

The tap_count and tap_dance_actions parameters should be the same as the similarly named parameters of the tapDanceAction function.

tapDanceAction(tap_dance_index, key_addr, tap_count, tap_dance_action)

The heart of the tap-dance plugin is the handler method. This is called every time any kind of tap-dance action is to be performed. See the How does it work? section for details about when and how this function is called.

The tap_dance_index and tap_count parameters help us choose which action to perform. The key_addr parameter tells us where the tap-dance key is on the keyboard.


Further reading

Starting from the example is the recommended way of getting started with the plugin.