Build your first automation in five minutes

The fastest way to understand AutoFlowRPA is to build something small that actually works. This one takes about five minutes and touches the three things every automation needs: reading input, doing work in a loop, and writing output.

What we are building. A spreadsheet holds a list of URLs. For each one we fetch the page, pull out the title, and write it back next to the URL.

1. Start with the canvas

Open AutoFlowRPA and create a new script. You get an empty canvas — an n8n-style node editor where each node is one command and the connections decide the order.

The command palette on the left is grouped by category: Excel, Web Browser, AI, Files, Video, and so on. Drag a command onto the canvas, or drop it directly on a connection to insert it between two existing steps.

2. Read the spreadsheet

Drag in Create Excel Workbook and point it at your file, then Get Range as DataTable to read the URL column. Give the output a variable name — say rows.

Every command that produces a value asks where to put it. That name is how the rest of the workflow refers to it.

3. Loop over the rows

Drop a Loop container on the canvas and feed it rows. Anything you place inside runs once per row, with the current row available as a variable.

Inside the loop, add HTTP Request. In the URL field, type:

{{{row.url}}}

Triple braces are how you reference a variable. They work in almost every field, and they understand indexing and dotted paths — {{{rows[0].url}}}, {{{response.data.title}}}. Arithmetic works too, through a safe evaluator, so {{{counter + 1}}} does what you expect without a scripting language.

4. Pull out what you need

Add Extract with Regex or Parse JSON depending on what came back, and store the result — title.

If you would rather drive a real browser than fetch raw HTML, swap the HTTP Request for Create Browser and Get Element Text, and pick the element visually instead of writing a selector by hand.

5. Write it back

Still inside the loop, add Set Cell: the sheet, the cell, and {{{title}}} as the value. After the loop, Save Workbook.

That is the whole workflow — about eight nodes.

6. Run it and watch

Press Run. The log below the canvas fills in step by step, with timing for each command, its output, and the variable values before and after. When something goes wrong you are not guessing: you can see the exact step, the exact input, and the exact error.

Where to go next

• Credentials — store API keys and OAuth secrets once, encrypted locally, and reference them from any script instead of pasting keys into fields.
• Triggers — run on a schedule, on a file appearing, or on a hotkey.
• Error handling — wrap risky steps in a Try container and decide what happens on failure instead of stopping the run.
• Sample workflows — the app ships with ready-made scripts for common tasks; opening one and taking it apart is a good second lesson.

There are 369 commands in total (339 on macOS). You do not need to learn them — the palette is searchable and every field has an inline hint. Build the small thing first; the rest is there when you need it.