OCR without mobile app

[Sincostan]

I have a family account and want to use OCR to create searchable PDF files. I want to use the following workflow:

1. Use Brother ADS 2800w to scan paper (German language) and send PDF file to a folder inside of my Dropbox (notice that I do not select "searchable PDF" but only "PDF" in the document scanner, since otherwise this would then firstly go to the Brother Cloud to use their OCR analysis)

2. Wait for Dropbox to autmatically (or can I trigger OCR?) apply the OCR analysis to my files, so that I can search for strings in those files let's say one day later.

Is this possible?

If yes, when will Dropbox analysis my files?

Or is the Dropbox OCR feature limited to the mobile app?

Best regards

[Mark]

It should happen quite quickly @Sincostan 

https://help.dropbox.com/files-folders/sort-preview/search-content 

[Sincostan]

@Mark, thank you for your reply. The documents are still not searchable. Do you use a comparable setup or do you use the mobile app?

[Sincostan]

Can someone from the Dropbox team look into this please?

[kls144]

Hi - also interested in an answer here. And if this varies between Plus and Professional accounts. Thank you

[Walter]

Hey @kls144 - thanks for joining the discussion here.

Could you clarify what exactly you'd like to accomplish?

Optical character recognition should work instantly on files you've scanned to Dropbox, both on the mobile app and the website. 

Also, may I ask if you're still having issues with this @Sincostan? 

[kls144]

Hi @Walter - thanks for your response. I would like to search over a large set of pre-existing handwritten/printed then scanned PDFs that were not scanned in via Dropbox's mobile document scanner. Is there a subscription plan that includes the functionality to search over PDFs generated outside of the Dropbox mobile app? 

I'm left unclear on what full text search encompasses as it relates to PDFs based on this list: https://help.dropbox.com/files-folders/sort-preview/search-content

The tech team did a great job of laying out the various PDF formats and how they can differ for searchability:  https://dropbox.tech/machine-learning/using-machine-learning-to-index-text-from-billions-of-images I'm left with the impression they addressed this in 2018. My use case is focused on what they call "Page has text, but only in the form of an image"

I'm currently a Plus prescriber. Thank you

[kls144]

Bumping for @Walter - thank you for the help!

[Walter]

Hi @kls144, thanks for the nudge on this!

I believe only scans generated by the built-in scanner can be searched using OCR at the moment while handwritten documents are also not detected at this time.

I'm double-checking this with the team as well and I'll let you know more when I hear back from them.

[kls144]

HI @Walter - did the engineering team ever confirm one way or another? Thank you