OCR - how it works?

Is it here anybody, who can explain me, how OCR works in Obsidian?

I paste this jpg picture and it works.

When I paste for example this one, it doesn´t.


I can see only this:


I tried it with different jpgs, but with no success.

There is no OCR in Obsidian as far as I can tell.

From what I see:

  1. First example is a remote image with alternative text. I assume what this was copied from the web.

Obsidian has the HTML to Markdown conversion when you paste something from the web. I also assume, what this alternative text stored on the website from which it was copied.

  1. Second example is when you copied an image file. It creates a file in your attachment folder with some placeholder file name. Then Obsidian put a link to that file and put file name as alternative text for image.

One nice thing here, if you use Dropbox to store your vault, they have OCR, so you get automatic OCR when searching your vault in the Dropbox app.

Makes sense. The 1st picture was taken from facebook. So maybe facebook scan pictures and add alt text to all pictures.

What it means to have vault in Dropbox?

Does it mean, that you have to be always online, so you will never use your desktop app offline? Because this OCR sounds very good.

Since Obsidian just operates on a standard folder of files, being ‘in Dropbox’, just means that you use Dropbox to sync that folder across your devices.

Is the result of this OCR visible to you, though, or is it simply used for searches within dropbox? At the moment I have tesseract running on this PC so that I can OCR stuff and get at the results — or you can photograph things on Android and run it through the OCR i MS Office Lens, and then paste the results. They are both annoying workarounds.

I do not believe it is visible for copying/pasting, just indexed for Dropbox searches.

There is an OCR plugin available since yesterday. Maybe this works for you?

1 Like

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.