Skip to main content
Keyboard commands let you send key presses, shortcuts (like cmd+c), and hold/release modifiers. These work at the application level, not the element level.

Overview

press

Send key combo (return, escape, cmd+s, shift+tab)

key-down

Hold a key or modifier down

key-up

Release a held key or modifier

Key Syntax

Single Keys

agent-desktop press return
agent-desktop press escape
agent-desktop press tab
agent-desktop press space
agent-desktop press delete
Arrow keys:
agent-desktop press up
agent-desktop press down
agent-desktop press left
agent-desktop press right
Function keys:
agent-desktop press f1
agent-desktop press f12

Key Combos with Modifiers

Use + to combine modifiers with keys:
agent-desktop press cmd+s
agent-desktop press cmd+c
agent-desktop press cmd+v
agent-desktop press cmd+z
agent-desktop press cmd+shift+z
agent-desktop press ctrl+a
agent-desktop press shift+tab
Modifiers: cmd, ctrl, alt, shift

Multi-Modifier Combos

agent-desktop press cmd+shift+z   # redo
agent-desktop press cmd+alt+esc   # force quit dialog
agent-desktop press ctrl+shift+eject  # lock screen

Common Patterns

Save Current Document

agent-desktop press cmd+s

Copy/Paste

agent-desktop press cmd+c
agent-desktop press cmd+v

Undo/Redo

agent-desktop press cmd+z        # undo
agent-desktop press cmd+shift+z  # redo
agent-desktop press tab          # next field
agent-desktop press shift+tab    # previous field
agent-desktop press return       # default button
agent-desktop press escape       # cancel

Hold Modifier for Multi-Selection

agent-desktop key-down shift
agent-desktop click @e3
agent-desktop click @e5
agent-desktop key-up shift
This selects multiple items in a list.
agent-desktop press cmd+space
agent-desktop wait 200
agent-desktop type @e1 "TextEdit"
agent-desktop press return

Examples

# Press single key
agent-desktop press return
agent-desktop press escape

# Press key combo
agent-desktop press cmd+s
agent-desktop press cmd+shift+z

# Hold and release modifier
agent-desktop key-down shift
agent-desktop click @e3
agent-desktop key-up shift

# Navigate with arrows
agent-desktop press down
agent-desktop press down
agent-desktop press return

# Function keys
agent-desktop press f5  # refresh in browser

Use Cases

Drive apps without mouse, using tab and arrow keys:
agent-desktop press tab
agent-desktop press tab
agent-desktop press return
Trigger application shortcuts:
agent-desktop press cmd+n      # new window
agent-desktop press cmd+w      # close window
agent-desktop press cmd+q      # quit app
Hold shift or cmd while clicking:
agent-desktop key-down cmd
agent-desktop click @e2
agent-desktop click @e5
agent-desktop key-up cmd
Use keyboard shortcuts for editing:
agent-desktop press cmd+a      # select all
agent-desktop press delete     # delete
agent-desktop press cmd+z      # undo

Supported Keys

Printable Keys

All alphanumeric characters: a-z, 0-9, and symbols.

Special Keys

  • return, enter
  • escape, esc
  • tab
  • space
  • delete, backspace
  • up, down, left, right
  • home, end
  • pageup, pagedown

Function Keys

f1 through f12

Modifiers

  • cmd (Command/⌘)
  • ctrl (Control)
  • alt (Option/⌥)
  • shift (Shift/⇧)

Error Handling

Common error codes:
  • INVALID_ARGS: Unrecognized key name or malformed combo
  • ACTION_FAILED: OS rejected the key event (rare)
All keyboard commands return structured JSON with error codes and recovery hints.

Build docs developers (and LLMs) love