I have tried quite a few OCR options over the last 6 years - and for me KAMUI is hands down the best solution. Especially if you are playing games on the PC. I run my Switch through the PC via OBS and it just works. Yeah, its $5 a month, but someone has to write it and maintain, so it seems a fair trade.
First, I recommend not mining until you learn at least 1,000 words from a premade list. You could add them to a custom deck, but it’s more imporant to get the basics down first.
For mining from games, you could look up the word you want to use in bunpro, and put it in a custom deck? I did that when I was mining from paper manga.
I’m currently mining from かがみの孤城 in anki using this method
Alright everyone, I’m here to update y’all on what I’ve been using for OCR.
I did try the one recommended by @Urazi , but unfortunately it didn’t work for me. I’m confident that I installed and used it correctly, but it just didn’t work.
I went to try Kamui OCR, but downloading it with this terrible internet speed where I am was just not for me. Also, it was a paid subscription after 500 scans, so I went for something else.
I wanted to set up GameSentenceMiner, but before I did, I said, “You know what? Let me try YomiNinja once again.” And surprisingly, it didn’t crash and it’s stable (mostly). But one complaint I have with it is that it takes about 2-3 seconds to scan the screen. I’m guessing it’s because of the internet speed, and yes, YomiNinja only works online.
On the plus side, YomiNinja is free and offers a really good UI with great extensions like 10ten and Yomitan pop-up dictionary.
Here’s a screenshot of how it works:
Thank you for testing those out, sad that dokidoki addict didn’t worked up for you but it’s still very early in development so maybe later you could try again?
It’s fully offline so it would be great in your case (online is only for gemini dictionnary).
Enjoy your game! 
Definitely try it again. I didn’t uninstall it, so I’ll figure out what the problem is. 
Wow! I didn’t know that! Thanks for mentioning it. 
which ocr model are you using?
I believe it’s Google lens. I double check tomorrow 
Yep, it’s Google Lens

