Read PDF

Extract text and metadata from PDF files

Overview

The Read PDF tool reads a PDF binary from an input item, extracts all text content page by page, and outputs document metadata (page count, info dictionary, XMP metadata) along with the extracted text. Supports password-protected/encrypted PDFs. The original binary is forwarded on the output item so downstream nodes can still access the PDF file. Uses pdfjs-dist (Mozilla PDF.js) for parsing.

Category: Core Nodes
Tool Name: read_pdf
Version: 1

Appearance: Icon: lucide-FileText | Color: #003355

Node Type

Action (Binary) — handles file/binary data operations

Input / Output

Direction	Port(s)
Input	Input
Output	Output, Error

Credentials

This tool does not require any credentials.

Parameters

Parameter	Type	Required	Default	Description
Binary Property	string	Yes	data	Name of the binary property containing the PDF file to extract text from. Names are case-sensitive — see the upstream node’s Binary Data panel for the exact names to use.
Password	string	No	—	Password to decrypt the PDF if it is encrypted/password-protected. Leave empty for non-encrypted PDFs.
Max Pages	number	No	0	Maximum number of pages to extract text from. Set to 0 to extract all pages.
Join Pages	boolean	No	true	Whether to join text from all pages into a single string separated by double newlines (true), or return an array of per-page text strings (false).
Keep Source	options	No	both	What data from the input item to keep on the output. “json” merges input JSON with extracted data, “binary” forwards binary, “both” does both.
				Options: json, binary, both
Max Concurrency	number	No	10	Maximum number of items to process concurrently.

Output Data

One output item per input item. The extracted data is merged into the item JSON at the top level:

{
  "numpages": 12,
  "numrender": 12,
  "info": {
    "Title": "Quarterly Report",
    "Author": "Dana Whitfield",
    "CreationDate": "D:20260731094500Z"
  },
  "metadata": {
    "dc:title": "Quarterly Report"
  },
  "text": "Page one text…\n\nPage two text…",
  "version": "4.2.67"
}

• numpages — the total number of pages in the document.
• numrender — how many pages were actually read. Lower than numpages when Max Pages limited the run or the text cap was reached.
• info — the PDF info dictionary (title, author, producer, creation date and so on), or null when the document has none.
• metadata — the document’s XMP metadata as a flat object, when present.
• text — the extracted text. A single string with pages separated by blank lines when Join Pages is on, or an array with one string per page when it is off.
• version — the version of the PDF engine that parsed the file.
• isScanned — added and set to true when the document has pages but yielded no text at all, which means the PDF is image-only and needs OCR.
• _warning — added when the result is partial or empty: either the extracted text hit its size cap, or nothing could be extracted.

Keep Source decides what survives from the input item:

Value	Output JSON	Output binary
both	Input item JSON merged with the extracted fields	The item’s binary is forwarded unchanged, including the PDF
json	Input item JSON merged with the extracted fields	The source PDF property is removed; any other binary properties are kept
binary	Only the extracted fields — the input JSON is dropped	The item’s binary is forwarded unchanged

Where a name collides, the extracted field wins over the input field.

Reference the text downstream by expression, e.g. {{ $json.text }}.

Usage Examples

• Extract all text from a PDF document
• Read a password-protected PDF
• Extract text from the first 5 pages of a large PDF
• Get PDF metadata like author, title, and page count
• Parse uploaded PDF and get per-page text as an array

Example Configuration

Extract everything from a PDF attached as data:

{
  "type": "read_pdf",
  "parameters": {
    "binaryPropertyName": "data",
    "maxPages": 0,
    "joinPages": true,
    "keepSource": "both"
  }
}

Read only the first five pages, one string per page:

{
  "type": "read_pdf",
  "parameters": {
    "binaryPropertyName": "data",
    "maxPages": 5,
    "joinPages": false,
    "keepSource": "json"
  }
}

Open an encrypted PDF:

{
  "type": "read_pdf",
  "parameters": {
    "binaryPropertyName": "invoice",
    "password": "letmein",
    "keepSource": "both",
    "maxConcurrency": 4
  }
}

Error Handling

Mode	Behavior
stop	Halts workflow on first error
continue	Skips failed items, passes successful ones through
errorPort	Routes failed items to Error output port

Tips

Extract text and metadata from PDF binary data — supports encrypted PDFs, page limiting, and per-page text output.

Behavior notes

• Scanned PDFs return no text. A document made of page images produces empty text, isScanned: true and a warning. Run it through an OCR tool instead of retrying here.
• Password errors are explicit. An encrypted PDF with no Password fails with “PDF is password-protected”; a wrong one fails with “PDF password is incorrect”. Neither is retried.
• Max Pages 0 is not unlimited. With no limit set, extraction stops after 1,000 pages; documents over 10,000 pages are refused outright. Set Max Pages when you only need the front matter — it is much faster on long documents.
• There are size caps. Files over 100 MB are rejected before parsing, and extraction stops once the collected text passes 50 MB, returning a partial result with _warning set.
• The PDF stays attached unless you set Keep Source to json, so you can extract text and still upload or archive the original file later in the workflow.