Transformers sit between the nodes that produce data and the nodes that consume it. They take a payload on their inlet, change its shape, and pass the result on. Nothing here touches the world outside the flow.
Three of them fan out rather than pass straight through. For Each and While each have a Loop outlet that runs everything downstream once per pass and a Done outlet that fires once at the end; Extract Text (PDF) has both shapes at once — Document for one answer about the whole file, Page for one per page.
Every Transformers node
Extract Text (OCR)Reads the text out of an image with Apple's Vision framework — entirely on this Mac, no network, no AI provider.Extract Text (PDF)Pulls the text out of any PDF — born-digital or scanned. Pages with no text layer go through on-device OCR automatically.For EachRuns everything downstream once per item in a list, then carries on.HTML TemplateWrite an HTML page once, with {{key}} filling in payload values, and emit it as text for whatever comes next.Image MetadataReads a photo's EXIF — capture date, camera, lens, exposure, GPS — and merges it onto the payload. On-device, no network.Map KeysReshapes the payload — each row builds a new key from a {{ }} value template.MathDoes arithmetic on payload values and puts the number back on the payload.TemplateBuilds a piece of text out of payload values, and puts it on the payload under a key you choose.WhileRepeats the loop body until a payload field says to stop — polling, retrying, paging.