The PDF steps work on PDF files between steps: read the text of a supplier's invoice so an AI step can pick out the total, merge a delivery note with its invoice, or make a PDF from HTML to attach to an email. Each one is its own step. No connection is needed.

Who can do this

Workspace Admins and Editors, on every plan.

The steps

Step What it does
Extract text from a PDF Reads the text of a PDF, page by page.
Merge PDFs Joins PDFs into one, in the order given.
Split a PDF Cuts a PDF into parts, one item each — a page each, or the ranges you give, such as 1-3, 4, 7-.
Make a PDF from HTML Turns HTML with values from earlier steps into a PDF of up to 50 pages, A4 or Letter.

Before you start

Except for Make a PDF from HTML, an earlier step must bring the PDF — a form's File field, Outlook · Get an attachment, or another step that gives a file. See Pass files between steps.

Steps

  1. Select + where the PDF is needed. In the step picker, open Utilities, select PDF — No connection needed — then the step. Or search PDF.
  2. Fill in the step's settings:
    • Extract text from a PDF — under File, select Add a file and drag in the PDF from Data from earlier steps.
    • Merge PDFs — under Files, select Add a file once for each PDF and drag each one in. PDFs from earlier steps, a line each, in order — the order of the rows is the order of the pages. In File name, keep merged.pdf or type your own, such as INV-1042.pdf.
    • Split a PDF — under File, drag in the PDF. In Pages, type the ranges, such as 1-3, 4, 7- (7- is page 7 to the end). A part for each range. Empty: one part per page.
    • Make a PDF from HTML — in HTML, write or paste the page, with values from earlier steps in {{ }}, such as <h1>Invoice {{ $json.invoice_no }}</h1>. Choose Page size — A4 (the default) or Letter — and Margin: In points, all round. 72 points are an inch — 20 unless you change it, up to 100. In File name, keep document.pdf or type your own.
  3. Select Run this step, then check Output. A new file shows with Download.

Output

Extract text from a PDF adds text (the whole text, the pages separated by a blank line), pages (a list, one text per page) and pageCount to the item.

Merge PDFs and Make a PDF from HTML add the new file to the item under file, with pageCount. Split a PDF gives one item per part, each with its file under file, pageCount, and pages — the range it holds, such as 1-3. A part's file is named after the original: INV-1042-p1-3.pdf.

The item's other fields are kept. Use the file in a later step by dragging it in from Data from earlier steps, or with {{ $json.file }}.

Good to know

  • A scanned PDF has no text to extract — it is a picture of a page. Extract text from a PDF gives empty text for those pages, without failing; an AI step that reads images can read them instead.

  • Make a PDF from HTML runs no script and fetches nothing from the internet. No scripts, nothing fetched: an image is a data: address or a file from an earlier step, styles go in a <style> block. For a file from an earlier step, write it as the whole of the image's address: <img src="{{ $json.file }}">. An image at a web address is left out. A <link> to a stylesheet and an @import are removed before the page is made.

  • The CSS Make a PDF from HTML supports is CSS 2.1-level styling:

    • fonts — family, size, weight and style — text colour and background colours;
    • borders, padding and margins on elements, widths and alignment;
    • tables, which are the way to lay out columns;
    • page-break-before and page-break-after, to start a new page.

    Flexbox and grid are not supported, and nor are @page rules or headers and footers repeated on every page. The page's margin is set by the step's Margin, not by CSS.

  • Fonts. Text is set in Noto Sans, which ships with Bizomate; font-family: Poppins gives Poppins, regular and bold. Any other font name falls back to Noto Sans. Noto Sans covers Latin, Greek and Cyrillic letters.

  • Make a PDF from HTML makes up to 50 pages. Longer HTML stops the step and says so; split the content over more than one step.

  • A PDF that needs a password to open cannot be merged or split, and its text cannot be read.

  • Files are kept with the run, like any file passed between steps.

  • Each step runs once per item, so each order or invoice gets its own PDF.

If something goes wrong

What the run says Why What to do
… can't be read as a PDF: … Extract text from a PDF got a file that is not a PDF, or one protected with a password. Check the earlier step's Output tab. Ask the sender for an unprotected copy.
… can't be opened as a PDF — it may be protected with a password, or not a PDF. (…) Merge PDFs or Split a PDF got a file that is not a PDF, or one protected with a password. Check the earlier step's Output tab. Ask the sender for an unprotected copy.
Pages “…” is not a page or range in this …-page PDF — write 1-3, 4, 7-. Pages names a page the PDF does not have, or is not written as ranges. Check the PDF's page count and correct the ranges.
Pages “…” runs backwards. A range ends before it starts, such as 5-2. Write the smaller page first: 2-5.
HTML is empty. Write the page, with values from earlier steps in {{ }}. HTML has nothing in it. Write the page.
The PDF would be … pages; one step makes up to 50. Split the content. The HTML makes more than 50 pages. Make the PDF in parts, then join them with Merge PDFs.
The HTML could not be made into a PDF: … The renderer could not lay out the HTML. Check the HTML for broken tags; simplify the styles.
Margin “…” is not a whole number from 0 to 100. Margin is not a whole number of points, or is over 100. Type a number from 0 to 100.
File is empty. Choose a file from an earlier step. / Files is empty. Choose files from earlier steps. No PDF was chosen. Drag in a PDF from Data from earlier steps.