LoopfourRead PDF documents
Pull the text and pages out of a PDF, such as a vendor bill or a statement, so the later steps in a workflow can read and use what it says.
The PDF step opens a PDF and gives later steps what is in it. Use it for bills, statements, remittance advice and contracts that arrive as PDFs.
- TriggerStatement arrives by emailThe PDF is saved to Files.
- StepPDF: Read textThe text of each page comes back.
- StepAI Agent: pick out the balances
- OutputBalances added to a table
Before you start
Have a sample PDF from an earlier step or Files. Check whether it contains selectable text or scanned pictures.
What it can do
Choose one under Operation:
- Read text: the text of the whole document, or of the pages you choose.
- Extract pages: a new, shorter PDF with only the pages you choose, such as the summary page of a long statement. It is saved to Files unless you choose otherwise.
- Get info: how many pages the PDF has.
Set it up
Add the PDF block
Drag PDF from Core Blocks onto the canvas, after the step that brings in the file, such as the Inbound Email trigger.
Choose the operation
Pick Read text, Extract pages or Get info under Operation.
Point it at the file
Under PDF File, point at the file from the earlier step. See Use results from earlier steps. If the PDF is on the web instead, put its address in Or a Link.
Choose pages, if you need to
Under Pages, list pages and ranges such as 1-5, 8. First Pages and Last Pages take a number of pages from the front or the back. Leave them empty for the whole document.
Good to know
- Read text reads the text stored in the PDF. A scanned page is a picture, so it has no text to read. For scans, give the file to an AI Agent step instead.
- Read text returns up to 500 pages, or 10 MB of text.
- The text comes back as written on the page. To pick out particular fields, such as the invoice total, follow it with an AI Agent step.
Check it worked
Run the workflow with a sample PDF. Open the run under Runs, choose the PDF step and read its Output.
If something goes wrong
If Read text returns no useful text, check whether the PDF is scanned. Use the AI route described above for scans. For a missing file or failed step, see Error messages explained.
Related
Read, extract and classify with AI
Learn what the AI Agent step does, such as pulling fields from a document or sorting items, and how to check its output before it is used.
Send an email or Slack message
Notify people from a workflow by sending an email or posting a Slack message, for example to tell AP that a bill is ready for review.