Overview

fileInput is the browser upload entry point for workflow files. It does not transform data. It simply returns the uploaded Blob or File objects so downstream nodes can process them.

Configuration / Settings

Setting	Type	Required	Details
Accepted Types	`pdf`, `image`, `word`, `excel`, `ppt`, `text`, `all`	No	Editor-side upload filter. It improves UX but does not replace downstream MIME validation.
Allow Multiple Files	boolean	No	Defaults to `true`. When enabled, downstream nodes receive an array of files.

Inputs Expected

This node does not require upstream input. At runtime it reads:

data.fileIds when files were stored in IndexedDB during a saved/editor session.
data.tempFiles when files are still in memory in the current browser session.

Variables Exported / Outputs

Output shape: Blob[]

The node completes with an array of files, even when only one file was uploaded.
Structured field access is not exposed directly because the output is binary data, not JSON-like records.
Filename-aware downstream nodes can still reuse the original File.name.

Common Use Cases

Start a PDF pipeline from manual uploads.
Feed a template file into pdfFormFill or wordFormFill.
Upload multiple PDFs before mergePdf, matchFiles, zip, or downloadOutput.

Edge Cases

If neither fileIds nor tempFiles exist, the node completes with an empty array.
Saved workflows intentionally clear transient upload data, so templates do not reopen with stale files attached.
File-type enforcement happens later. A downstream conversion node can still reject a file that the upload picker allowed.