The task editor

The task editor is where you build, check, run and debug a task: a command palette on the left, the task in the middle as blocks or as code, and the properties of the selected command on the right.

#Tasks and folders

Open Tasks to see your automations. The switch at the top chooses the workspace:

See Versions and workspaces for pull, push and clone.

• FOLDERS on the left organize tasks. The New folder button next to the heading creates a folder inside the selected one; a folder's menu has New subfolder, Rename and Delete (only an empty folder can be deleted). All tasks shows every task.
• New task asks for a Name, a Description and a Folder. New tasks start in your local workspace and open in the editor.
• A task's menu (⋯) has the workspace actions, Edit details and Delete. Edit details also holds the task's performance settings and its Self-healing choice (When an element is not found); see Finding elements.
• Click a task to open it in the editor.

Note: Only users with the Edit tasks permission (Developer by default) create and change tasks. A public version is read-only in the editor; pull the task to change it.

#Editor layout

The panels can be resized by dragging their borders. The line under the center switch shows the main shortcuts for the current view.

#Toolbar

• Workspace tags next to the name tell you what you are looking at: Public · v4 and Read-only, Local · pulled from v4, Local · never pushed, or Local clone of v4. Pulled by shows who holds the lock on a public task.
• Unsaved changes appears after an edit; Not pushed means your saved local copy differs from the public version.
• No problems, Problems or To complete summarizes validation. Click it for the list; click an entry to jump to the block (or to the code, for syntax errors).
• Undo and Redo (Ctrl+Z, Ctrl+Y or Ctrl+Shift+Z).
• Record records what you do in Chrome and in Windows programs; see Recording. AI Task Generator writes commands from a description; see AI features.
• History opens the version history, the compare button next to it compares versions; see Versions and workspaces.
• The documentation button next to them opens the task's documentation (technical description, developer notes, generated summary, PDF); see Task documentation.
• Run starts the task on an agent, Debug starts it with the debugger (Debug tasks permission); see Running tasks and Debugging.
• Save (Ctrl+S) writes your local copy. Push publishes it as the next public version.

Note: Run and Debug are disabled while you have unsaved changes: agents run the saved local copy (or the public version), never what is only in your browser.

#Command palette

The COMMANDS tab lists every command in groups (Variables, Condition, Loop, Flow Control, Error Handling, Wait & Delay, Messages & Logs, Web, Desktop, Data, Files & Folders, Email & API, Scripts & Programs, AI). One group is open at a time; hover a command to read what it does.

• Search commands filters by name and description and opens every group with a match.
• Click a command to add it after the selected block (into a selected section, or at the end when nothing is selected).
• Drag a command to drop it exactly where you want; a blue line shows where it lands.
• The sections Else if, Else, Catch and Finally are listed under If and Try. Click one with an If or Try block selected, or drag it onto the block.

In a read-only task the palette is replaced by a note: pull the task to add or change commands.

#Add a command with the keyboard (Ctrl+K)

The quickest way to build a task: press Ctrl+K (or Add command above the blocks), type what the task should do and press Enter.

• It finds commands and their actions by name, description and everyday words: excel read, click, type text, send email, for each, wait, approve all find what they mean. Every word must match; the best matches come first.
• An action found this way is already chosen in the new block, e.g. Excel · Read adds excel --action:read.
• Enter adds after the selected block and opens its properties. Ctrl+Enter adds and keeps the box open, to add several blocks in a row. ↑ ↓ choose, Esc closes.
• Before you type, the box lists the commands you used last, then the most common ones.
• Sections (Else if, Else, Catch, Finally) are found too and go into the selected If or Try block.

#Blocks view

Every command is a card with its icon, name and a one-line summary of its parameters. Block commands (If, Loop, Try, Step) have a colored edge, indent their contents and end with an End … line. The number left of each row is its line in the task text: logs, errors and the debugger refer to that number.

• Select a block by clicking it. The arrow left of a block command folds it; a folded block shows how many commands it hides (and N inside when some of them have problems).
• Move blocks by dragging them.
• The ⋯ menu of a block has Run from here, Debug from here, Add breakpoint, Copy, Cut, Duplicate, Disable and Delete; on an If block also Add else if and Add else, on a Try block Add catch and Add finally.
• The End if and End try lines offer + Else if, + Else, + Catch and + Finally for the sections the block does not have yet.

#Selecting several blocks

With several blocks selected, the properties panel shows N blocks selected with Copy, Cut, Duplicate, Delete and Disable/Enable, and dragging one of them moves them all. A nested block always moves and copies together with its parent.

#Copy and paste as text

Copy and cut put the selected blocks on the clipboard as task text. You can paste them into another task, into the Code view, into a chat or a ticket, and paste task text back: Ctrl+V in the blocks view inserts the commands after the selection. If the clipboard holds no valid commands, the editor says so and changes nothing.

#Disabling commands

Ctrl+/ (or Disable in the block menu or the properties panel) switches the selected commands off. A disabled command stays in the task, greyed out with a Disabled badge, but does not run and is not validated; a disabled block takes everything inside it along. Press Ctrl+/ again to enable it. In the task text a disabled command carries --disabled:true.

log --message:"Only while testing" --disabled:true

#Steps

A Step groups commands under a name, such as "Log in to the portal". It does nothing itself; the commands inside run in order. Use steps to structure long tasks: they fold like other blocks, and Run from here works on a step and on the commands inside it.

step --description:"Log in to the portal"
    browser --action:open --url:"https://portal.example.com" --output:portal
    capture --browser:$portal$ --action:settext --dompath:"#user" --value:$login.Username$
endstep

#Code view

Code shows the same task as text in the task language, with syntax highlighting, suggestions and error markers. Changes in either view appear in the other.

• Suggestions appear as you type (also after --, : and $); Ctrl+Space opens them on demand. Hover a word for help.
• F2 opens the variable list at the cursor.
• F9, or a click in the margin left of the line numbers, sets or removes a breakpoint on the command of that line.

When the text has a syntax error, the blocks view shows the last valid version with a warning and a Fix in code button, and you cannot save until the error is fixed.

#Properties panel

Select a block to edit it in PROPERTIES. The header shows the command's name and description. The ? button at its right opens the command's documentation in a popup: what it does, every action with its parameters (the action the block uses is marked used here) and examples, as in the command reference. For a section such as Catch or Else it shows the block it belongs to.

• Commands with several forms start with an Action or Type list; the fields below follow your choice.
• Required fields come first and are marked with *. Fields you rarely need are under Optional settings (N). The info icon next to a field explains it.
• If and Try blocks have a Sections area to add else if, else, catch and finally. A section (such as Catch) has its own properties and a Remove … button.
• Duplicate, Disable and Delete at the bottom act on the selected block.
• Fields that point at something offer a button: Capture (a web or desktop element, or an anchor), Choose (an open window), New (a window variable), Browse (a file or folder on the agent machine), Choose task (for Run task).

#Object properties

Commands that act on a web or desktop element have an Object properties table: the screenshot taken when you captured the element, with the element marked, its Object type (Text box, Button, Check box, Drop-down list, Link…) and technology, and every property read from it. The element is found by the checked properties, all of which must match.

• Check or uncheck properties; the selector above is rebuilt from them.
• Each checked value must be equal to the element's. For the parts that change, type a *: it makes the value a wildcard. Invoice* starts with Invoice, *total* contains total, *.pdf ends with .pdf. $variables$ work too.
• Tasks made before this kept their other comparisons (Contains, Starts with…), shown as a tag next to the property; close the tag to compare as equal.
• For desktop elements, choose the technology next to the Capture button before capturing: Auto (UI Automation 3, and UI Automation 2 where it shows more of the element), UI Automation 3 (modern applications) or UI Automation 2 (older Win32 and WinForms applications). The task finds the element with the same technology when it runs. See Desktop automation.
• For web elements, type another attribute name (such as data-id) and click Add.
• If you edit the selector by hand, the table is no longer used; Use properties rebuilds the selector from the table.

See Finding elements for fallbacks, anchors and self-healing.

#Run task mappings

For Run task, choose the task with Choose task. Mappings then lists its Inputs and Outputs. Auto map fills empty inputs from same-named variables and sends outputs to same-named variables. See Variables.

#Inserting variables

Every field that accepts a variable has a $ button (or press F2 in the field) that opens Insert variable:

• Type to search, use ↑ and ↓ to move and Enter to insert.
• For a list, dictionary, record or data table, press → (or click it) to pick a part: One item, One key / field, One cell, One row (record), or Number of items / Number of rows. For text you can insert Its length. The reference is built for you, such as $items[0]$, $order.Total$ or $table[2].Name$.
• New variable declares a variable and inserts it. When nothing matches your search, Create offers the search text as the name.

Typing $ in a field also lists the variables; Enter or Tab inserts one.

#Variables panel

The VARIABLES tab lists every variable of the task, grouped as INPUTS, OUTPUTS, VARIABLES, FROM COMMANDS (loop variables and command outputs) and BUILT-IN ($System$, see System variables), with its type and initial value.

• Click a variable to select the command that declares it; double-click to edit it.
• Renaming a variable changes it everywhere in the task.
• Add variable (the + button) opens a dialog with Variable, Input or Output, Name, Type and the initial (or default) value. The declaration is added at the top of the task.

#Validation

The editor checks the task as you work. Each block shows its state:

You cannot save while there are errors or empty required fields; the toolbar's Problems or To complete list takes you to each one.

Tip: Apply on a healed block writes the properties of the element that was used into the object properties table. Save the task (and push it) to keep the change.

#Keyboard shortcuts