> ## Documentation Index
> Fetch the complete documentation index at: https://tbd-6fc993ce-hypeship-mcp-screenshot-tool-docs.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# computer_action

> Mouse, keyboard, and clipboard input at screen coordinates

Drive a browser session with raw mouse and keyboard input at screen coordinates. Pass a single action for simple operations, or pass multiple actions to batch them into one request for lower latency.

<Warning>Prefer [`execute_playwright_code`](/reference/mcp-server/tools/execute-playwright-code) for anything a selector can reach. It's faster, deterministic, and doesn't depend on the model's ability to locate a target in an image. Reach for `computer_action` only when there's no selector to target: canvas apps, embedded PDFs, native dialogs, and drag interactions.</Warning>

Coordinates come from a [`screenshot`](/reference/mcp-server/tools/screenshot) call, and are only as accurate as the calling model's pixel grounding — models vary widely here. Take a screenshot after acting to confirm the result.

## Parameters

| Parameter    | Description                                   |
| ------------ | --------------------------------------------- |
| `session_id` | Browser session ID. Required.                 |
| `actions`    | Ordered list of actions to perform. Required. |

## Action types

| Type                 | Description                                                                    |
| -------------------- | ------------------------------------------------------------------------------ |
| `click_mouse`        | Click at `x`, `y`. Supports `button`, `click_type`, `num_clicks`, `hold_keys`. |
| `move_mouse`         | Move the cursor to `x`, `y`.                                                   |
| `type_text`          | Type `text`, with optional inter-key `delay`.                                  |
| `press_key`          | Press `keys` (X11 keysym names or combos like `Ctrl+t`, `Return`).             |
| `scroll`             | Scroll at `x`, `y` by `delta_x`/`delta_y` (positive = right/down).             |
| `drag_mouse`         | Drag along a `path` of `[x, y]` points.                                        |
| `set_cursor`         | Show or hide the cursor (`hidden`).                                            |
| `sleep`              | Wait `duration_ms` between steps when the page needs time to react.            |
| `write_clipboard`    | Write `text` to the clipboard.                                                 |
| `read_clipboard`     | Return the current clipboard contents.                                         |
| `get_mouse_position` | Return the current cursor position.                                            |

<Note>`read_clipboard` and `get_mouse_position` return data, so they must come last. To capture the page, call the separate [`screenshot`](/reference/mcp-server/tools/screenshot) tool.</Note>

## Example

```json theme={null}
{
  "session_id": "browser_abc123",
  "actions": [
    { "type": "click_mouse", "click_mouse": { "x": 420, "y": 300 } },
    { "type": "type_text", "type_text": { "text": "kernel browsers" } },
    { "type": "press_key", "press_key": { "keys": ["Return"] } },
    { "type": "sleep", "sleep": { "duration_ms": 1000 } }
  ]
}
```

Then call `screenshot` to see the result.