Module 0codeKit PDF to OCR input issue

[Razvan]

Hi.

I am trying to build a scenario where the automation would take a scanned .pdf file, get the text using OCR, send the text to GPT, get a JSON and insert data into airtable. Below you can find a screenshot of the desired workflow:

My current problem is that I get errors when I try to send the .pdf URL to 0codeKit for getting the text using OCR. It seems that it expects some kind of .pdf / link as input and I am not providing the right thing. I also tried postman and I have the same issues. The dummy pdf works, but what I give as input does not work.

I’ve tried with both google drive and OneDrive.
Link to google drive pdf: * * * https://drive.google.com/file/d/1zKhbX3bwcw_aK14i3TrUWHOicBkDdSQ0

Link to OneDrive pdf: Microsoft OneDrive

The error I get using Google Drive pdf link: [500] Internal Server Error. The server met an unexpected condition

The error I get using OneDrive pdf link: Invalid PDF url provided.

Bundle from google drive:

What am I missing?

Thank you,
Razvan

[samliew]

Welcome to the Make community!

I think the permission should be reader/everyone.

reader/user only grants access to manually added users, which prevents 0CodeKit from accessing it.

[Razvan]

Hi @samliew

Both links can be accessed by anyone. I even tried opening them in incognito and they can be viewed.

I just checked the Google Drive permission and it’s set to viewer and everyone with the link:

[Sjoerd_Tiemensma]

I had a similar issue with Onedrive, where I couldn’t get a proper usable download link. One way of getting around this is to use the download a file module instead, then use 0codekit for temporary upload and use that URL