Skip to main content
Use this guide to choose the action that matches the step you want to automate. Each description focuses on what the action does for you and how to keep runs reliable once they are in production. Open a new page in the sandboxed browser. Provide the full URL so Komos can resolve redirects and log the final page in run history. Use separate Navigate actions for meaningful jumps between sites or app sections.

Click (CLICK)

Trigger buttons, links, and other clickable elements. Combine it with Wait For when the click changes the view, and use the selector tools in the builder to confirm you are targeting the right element.

Input (INPUT)

Fill text boxes, search bars, or form fields. Supply the desired text directly or pull it from task inputs. If the field is sensitive, connect it to stored credentials instead of typing secrets into the node.

Scroll (SCROLL)

Move the page to reveal content further down the screen. Add multiple scrolls back-to-back for long lists or infinite-scroll feeds.

Key Press (KEY_PRESS)

Send keyboard shortcuts such as Enter, Escape, or arrow keys. Helpful for confirming modals or advancing components that do not expose clear buttons.

Timing controls

Wait (WAIT)

Pause for a fixed amount of time. Use this for predictable delays, like waiting half a second for an animation to finish. For anything dynamic, reach for Wait For instead.

Wait For (WAIT_FOR)

Tell Komos what should appear before the run continues. Provide a snippet of text or a selector that signals the page is ready. This keeps automations fast and stable when network speeds vary.

Capture and transform data

Extract Data (EXTRACT_DATA)

Capture structured information from the current page. Add a field for every value you need, give it a clear name, and describe what should be collected so the recorder or planner can produce reliable selectors. Mark a field as a list when you expect multiple items.

Process Data (PROCESS_DATA)

Transform captured variables with a short prompt. Describe the output you want (summary, list, flag) and Komos will store the result for later nodes or task outputs.

File Output (FILE_OUTPUT)

Save information gathered during the run into a downloadable file. Use text mode for human-readable reports or JSON mode to pass structured data downstream.

Files and downloads

Browser Upload (BROWSER_UPLOAD)

Attach a file inside the sandboxed browser. Point the action at the upload control and map it to a file variable you collected earlier in the run. variable you collected earlier in the run.

Browser Download (BROWSER_DOWNLOAD)

Click a download button in the browser and keep the saved file with the run so you can retrieve it later.

File Download (FILE_DOWNLOAD)

Fetch a file directly from a URL without clicking through the UI. Perfect for APIs or export links that can be accessed with a simple GET request.

Communication and access

Login (LOGIN)

Apply saved credentials at the right moment in your flow. The action keeps passwords secure, tracks usage, and respects any rotation you perform inside Komos.

Send Email (EMAIL_SEND)

Email teammates or downstream systems straight from a run. Personalize the subject and body with variables you collected earlier in the automation.

Logic tools

If / Else (IF_ELSE)

Split the run into two paths when a decision is required. Describe the condition in plain language; Komos evaluates it and sends the run down the matching branch.

Loop (LOOP)

Repeat a set of steps for every item in a list (orders, rows, cards). Inside the loop you can refer to the current item and its index, then rejoin the main flow once the list is complete.

Tips for smoother runs

  • Pair actions with a Wait For that confirms the new page state before extracting data.
  • Preview runs after editing nodes to confirm selectors still match the latest UI.
  • Keep node descriptions concise and action oriented so collaborators immediately understand what the automation is doing.