Accessible via: Main Menu -> Transformations

A transformation turns one file format into another: extract structured data out of a PDF, normalise an inconsistent CSV, or run any custom conversion your data needs before something else can use it.

Transformations exist because incoming files rarely arrive in the shape your automation wants. Rather than teaching every workflow to cope with every export format, you normalise once at the edge.

Creating a transformation

1. Open Transformations in the left rail.
2. Click New Transformation.
3. Define what comes in, what should come out, and the steps between.

Where transformations fit

• Before a table. Normalise an export, then write clean rows into a table — instead of importing a file whose columns move between exports.
• Before a knowledge base. Get text out of documents so retrieval has something to index.
• Inside a workflow. Run a transformation as a step, so a file that arrives by email or webhook is usable by the time the next node sees it.

Transformations vs ETL pipelines

	Use it for
Transformations	One file in, one shape out. Format-level work.
Knowledge base ETLs	Keeping a whole source — a folder, a SharePoint site — continuously indexed.

If you’re converting a file, it’s a transformation. If you’re keeping a corpus in sync, it’s an ETL.