Practical automation patterns, explained step by step.
Learn how an outcome becomes a visual workflow. Each guide defines the inputs, outputs, control flow, failure cases, and the exact points where AI helps—or is unnecessary.
Automate web page extraction to Excel.
A complete example covering element capture, similar-item extraction, pagination, validation, spreadsheet output, and recovery when a page changes.
Collect paginated web data and save it to Excel
Build an editable workflow that reads repeated page elements, follows pagination, validates results, and writes a clean table.
Open the complete guide → REPAIR PATTERNDiagnose an element-not-found failure
Understand how the workflow, selector, current page state, runtime log, and error stack combine into a useful diagnosis.
Learn the debugging approach → DESIGN PATTERNBuild once, run repeatedly
Separate AI-assisted design from deterministic runtime execution so repeat work does not require a fresh model decision.
Read the design guide →A useful template must explain more than the happy path.
A workflow page should not be a thin list of commands. It should make the task reproducible by stating what data is expected, how elements are selected, when a loop stops, where output is stored, and what happens when assumptions fail.
The library begins with fewer, fully explained workflows. Additional guides will be added only when they include real structure, boundary cases, and a result that users can evaluate.
Describe the result. Let AI draft the visual flow.
Open the generated workflow in the editor, change any command or parameter, and keep the approved version for repeat execution.