> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tapkit.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Gestures

> Touch gesture methods - tap, swipe, drag, and pinch

Gesture methods accept either coordinates `(x, y)` or natural language descriptions. When you pass a string, TapKit uses vision AI to find and interact with the described element.

## Taps

```python theme={null}
# Coordinates
phone.tap((100, 200))
phone.double_tap((100, 200))
phone.hold((100, 200), duration_ms=1000)

# Natural language (vision AI)
phone.tap("the blue Submit button")
phone.double_tap("the photo to zoom in")
phone.hold("the app icon to delete")
```

| Method                           | Parameters                            | Description |
| -------------------------------- | ------------------------------------- | ----------- |
| `tap(target)`                    | coordinates or description            | Single tap  |
| `double_tap(target)`             | coordinates or description            | Double tap  |
| `hold(target, duration_ms=1000)` | coordinates or description, hold time | Long press  |

## Swipes

```python theme={null}
# Coordinates
phone.flick((500, 1000), direction="up")
phone.pan((500, 1000), direction="up", duration_ms=500)

# Natural language (vision AI) - flick only
phone.flick("the message list", "down")
phone.flick("the photo gallery", "left")
```

| Method                                   | Parameters                            | Selector Support | Description         |
| ---------------------------------------- | ------------------------------------- | ---------------- | ------------------- |
| `flick(target, direction)`               | coordinates or description, direction | Yes              | Quick swipe gesture |
| `pan(point, direction, duration_ms=500)` | coordinates only                      | No               | Controlled scroll   |

Directions: `"up"`, `"down"`, `"left"`, `"right"`

## Drag

```python theme={null}
# Coordinates
phone.drag((100, 200), (300, 400))
phone.hold_and_drag((100, 200), (300, 400), hold_duration_ms=500)

# Natural language (vision AI) - drag only
phone.drag("the slider handle", "the right end of the slider")
phone.drag("the file icon", "the trash folder")
```

| Method                                                      | Parameters                            | Selector Support | Description          |
| ----------------------------------------------------------- | ------------------------------------- | ---------------- | -------------------- |
| `drag(from_target, to_target)`                              | both coordinates or both descriptions | Yes              | Drag between points  |
| `hold_and_drag(from_point, to_point, hold_duration_ms=500)` | coordinates only                      | No               | Hold before dragging |

<Note>
  For `drag()` with selectors, both arguments must be strings. You cannot mix coordinates and descriptions.
</Note>

## Pinch

```python theme={null}
# Coordinates
phone.pinch((500, 500), action="pinch_out", duration_ms=1000)

# Natural language (vision AI)
phone.pinch("the map", "pinch_out")
phone.pinch("the photo", "pinch_in", duration_ms=1500)
```

| Method                                    | Parameters                                           | Description               |
| ----------------------------------------- | ---------------------------------------------------- | ------------------------- |
| `pinch(target, action, duration_ms=1000)` | coordinates or description, action, gesture duration | Pinch/zoom/rotate gesture |

| Action       | Effect                      |
| ------------ | --------------------------- |
| `pinch_out`  | Zoom in (fingers apart)     |
| `pinch_in`   | Zoom out (fingers together) |
| `rotate_cw`  | Rotate clockwise            |
| `rotate_ccw` | Rotate counter-clockwise    |

## Tips for Selectors

* Be specific: `"the blue Submit button"` works better than `"button"`
* Include visual details: color, position, text content
* Works best with clearly visible, distinct UI elements