Skip to main content
Returns the readable text of a PDF as JSON.

​
Inputs

Extraction Parameters accepts:
readingOrder helps on multi-column layouts, but on dense designs it can interleave sentences from adjacent columns. Unordered extraction returns the same output every time for the same file, so leave it off when you need reproducible results.
There is no Accept input — extractions always return JSON.

​
Outputs

A single result property holding the document’s text:
In the designer this surfaces as a result token you can drop into later steps.
There is no per-page breakdown in the response — result is one string for the whole request. To process pages separately, call the action once per page with pageIndices set to that page.

​
Example flow

Extract the text of an emailed invoice and post it to Teams for review.
1

When a new email arrives (V3)

Outlook, filtered to a shared mailbox and Only with attachments set to Yes.
2

Apply to each — Attachments

Iterate the trigger’s Attachments.
3

Extract all Text from PDF

Nitro. PDF File → the attachment’s Content. Leave Extraction Parameters empty to read the whole document.
4

Post message in a chat or channel

Teams. Use the Nitro step’s result output in the message body.
To pull a specific value rather than the whole document, use Extract searched Text from PDF, Extract PDF Form Data or Extract Invoice Data from PDF instead — all return structured results that are easier to branch on than free text.