input.performActions command

Syntax

{
  "method": "input.performActions",
  "params": {
    "context": "<contextId>",
    "actions": [
      {
        "type": "<outerType>",
        "id": "<sourceId>",
        "actions": [
          {
            "type": "<innerType>",
            ...
          }
        ]
      }
    ]
  }
}

Parameters

The params field contains:

context

A string that contains the ID (UUID) of the context in which to perform the actions. Context IDs are returned by commands such as browsingContext.getTree.

actions

An array of objects, each representing an input source and the actions to perform for that source. Each such object represents the outer actions object, which in turn, contains an outer type (input source type can be "key", "pointer", or "wheel") and an inner actions array. Each object in the inner actions array has its own inner type and additional fields that depend on it.

All input sources are processed in parallel. In each tick (step), every input source performs one action simultaneously or does nothing if assigned a "pause" action. This allows combining input sources, for example, holding Shift while clicking.

Each outer actions object has the following fields:

id

A string that uniquely identifies this input source within the action sequence, for example, "mouse1" or "keyboard1".

type

A string (the outer type) that identifies the type of input source. Accepted values are "none", "key", "pointer", and "wheel".

actions

An array of objects (the inner actions), each representing an action for the input source specified in the outer type field.

Each inner actions object has an inner type field that specifies the operation to perform and additional fields that depend on it. The inner type accepts the following values:

• "pause": Waits for the given duration before the next step.
• "keyDown": Simulates pressing a key.
• "keyUp": Simulates releasing a key.
• "pointerDown": Simulates pressing a pointer button.
• "pointerUp": Simulates releasing a pointer button.
• "pointerMove": Simulates moving the pointer.
• "scroll": Simulates a mouse wheel scroll.

The following table shows, for each outer type value, the valid values for the inner type:

Outer type values	Accepted inner type values
"none"	"pause"
"key"	"pause", "keyDown", "keyUp"
"pointer"	"pause", "pointerDown", "pointerUp", "pointerMove"
"wheel"	"pause", "scroll"

The following table shows, for each inner type value, the fields available in the inner actions object:

Inner type values	Fields available in the inner actions object
"pause"	duration
"keyDown", "keyUp"	value
"pointerDown"	button, pointer properties
"pointerUp"	button
"pointerMove"	x, y, duration, origin, pointer properties
"scroll"	x, y, deltaX, deltaY, duration, origin

The outer actions object also supports the following field:

parameters Optional

An object with a pointerType field that specifies the pointer device type. Accepted values are "mouse" (default), "pen", or "touch". This field is valid only when the outer type is "pointer".

The following fields are available in each inner actions object, depending on the inner type:

button

A non-negative integer that identifies the pointer button (0 = primary, 1 = middle, 2 = secondary). Specify this when the inner type field value is "pointerDown" or "pointerUp".

deltaX

An integer that specifies the horizontal scroll delta in CSS pixels. Specify this when the inner type field value is "scroll".

deltaY

An integer that specifies the vertical scroll delta in CSS pixels. Specify this when the inner type field value is "scroll".

duration Optional

A non-negative integer that specifies the time in milliseconds over which the action is performed. Specify this when the inner type field value is "pause", "pointerMove", or "scroll". For "pointerMove" and "scroll", the overall movement occurs as a series of small movements over this period at a browser-defined rate (for example, one step per animation frame). When multiple outer actions objects run in parallel, the tick lasts as long as the longest duration value in that tick.

origin Optional

A string or an object that specifies the origin for the move or scroll. Specify this when the inner type field value is "pointerMove" or "scroll".

If origin is a string, accepted values are:

• "viewport": Indicates that the x and y coordinates are relative to the viewport's top-left corner. Use this for absolute positioning within the page. This is the default value for "scroll" if origin is omitted.
• "pointer": Indicates that the x and y coordinates are relative to the current pointer position. Use this for relative moves from where the pointer currently is.

If origin is an object, include the following fields:

type

A string set to "element".

element

An object containing the ID that uniquely identifies the DOM element to use as the origin. The ID is returned by the browser when you locate the element using browsingContext.locateNodes, script.evaluate, or script.callFunction.

value

A string that contains the key value. Specify this when the inner type field value is "keyDown" or "keyUp". For special keys such as Shift or Enter, use the Unicode code points defined in the WebDriver keyboard actions table (for example, "\uE008" for the Shift key). For printable characters, use the character directly (for example, "a").

Pointer properties

The following fields are part of the inner actions object and describe the physical characteristics of the pointer device, such as a mouse, stylus, or touchscreen. Specify these when the inner type is "pointerDown" or "pointerMove".

width Optional

A non-negative integer that specifies the width, in CSS pixels, of the pointer contact area. See PointerEvent.width.

height Optional

A non-negative integer that specifies the height, in CSS pixels, of the pointer contact area. See PointerEvent.height.

pressure Optional

A float that specifies the normalized pressure of the pointer in the range 0.0–1.0. See PointerEvent.pressure.

tangentialPressure Optional

A float that specifies the normalized tangential pressure in the range -1.0–1.0. See PointerEvent.tangentialPressure.

twist Optional

An integer that specifies the clockwise rotation, in degrees, of the pointer in the range 0–359. See PointerEvent.twist.

altitudeAngle Optional

A float that specifies the altitude angle, in radians, of the pointer in the range 0.0–π/2. See PointerEvent.altitudeAngle.

azimuthAngle Optional

A float that specifies the azimuth angle, in radians, of the pointer in the range 0.0–2π. See PointerEvent.azimuthAngle.

x

A number (for "pointerMove") or an integer (for "scroll") that specifies the x-coordinate. Specify this when the inner type field value is "pointerMove" or "scroll".

y

A number (for "pointerMove") or an integer (for "scroll") that specifies the y-coordinate. Specify this when the inner type field value is "pointerMove" or "scroll".

Return value

The result field in the response is an empty object ({}).

Errors

invalid argument

The action sequence is malformed; for example, if a required field is missing, a field value is of the wrong type, or an outer type value is not "none", "key", "pointer", or "wheel".

no such frame

No context with the given context ID is found.

Examples

Holding Shift while clicking an element

Consider a scenario where you want to hold the Shift key while clicking an element, for example to extend a text selection.

With a WebDriver BiDi connection and an active session, get the context ID using browsingContext.getTree and the element identifier using browsingContext.locateNodes. Send the following message with two outer actions objects — a "key" source and a "pointer" source — each with an outer type and an inner actions array, running in parallel across the following three ticks:

• Tick 1: The keyboard presses Shift while the pointer moves to the element. Since the duration of pointerMove is specified as 300, the tick lasts 300 ms, which is the longest duration in this tick.
• Tick 2: The keyboard pauses while the pointer button is pressed (pointerDown). This tick lasts for 0 ms.
• Tick 3: The Shift key is released (keyUp) and the pointer button is released (pointerUp) simultaneously. This tick also lasts for 0 ms.

{
  "id": 1,
  "method": "input.performActions",
  "params": {
    "context": "5f07e3ca-ecac-465e-b9ef-49000c196ecf",
    "actions": [
      {
        "type": "key",
        "id": "keyboard1",
        "actions": [
          {
            "type": "keyDown", // Tick 1: Shift key down (0 ms)
            "value": "\uE008"
          },
          {
            "type": "pause" // Tick 2: Keyboard pause (0 ms)
          },
          {
            "type": "keyUp", // Tick 3: Shift key up (0 ms)
            "value": "\uE008"
          }
        ]
      },
      {
        "type": "pointer",
        "id": "mouse1",
        "parameters": {
          "pointerType": "mouse"
        },
        "actions": [
          {
            "type": "pointerMove", // Tick 1: Move to element (300 ms)
            "x": 0,
            "y": 0,
            "duration": 300,
            "origin": {
              "type": "element",
              "element": {
                "sharedId": "3be28343-afd3-4dea-a2b6-a863fbbb80e1"
              }
            }
          },
          {
            "type": "pointerDown", // Tick 2: Press button (0 ms)
            "button": 0
          },
          {
            "type": "pointerUp", // Tick 3: Release button (0 ms)
            "button": 0
          }
        ]
      }
    ]
  }
}

The browser responds as follows:

{
  "id": 1,
  "type": "success",
  "result": {}
}

Scrolling the page

Consider a scenario where you want to simulate scrolling a page down.

With a WebDriver BiDi connection and an active session, get the context ID using browsingContext.getTree. Send the following message that scrolls from the top-left of the viewport (x: 0, y: 0) by 300 CSS pixels downward (deltaY: 300) with no horizontal scrolling (deltaX: 0).

{
  "id": 2,
  "method": "input.performActions",
  "params": {
    "context": "5f07e3ca-ecac-465e-b9ef-49000c196ecf",
    "actions": [
      {
        "type": "wheel",
        "id": "wheel1",
        "actions": [
          {
            "type": "scroll",
            "x": 0,
            "y": 0,
            "deltaX": 0,
            "deltaY": 300
          }
        ]
      }
    ]
  }
}

The browser responds as follows:

{
  "id": 2,
  "type": "success",
  "result": {}
}

Specifications

Browser compatibility

See also

• input.releaseActions command
• input.setFiles command
• input.fileDialogOpened event