A suggestion and asking for advice

General
1

Most of the texts I read are scanned, and contain some OCR mistakes. Is there a way to get these fixed automatically? I tried asking ChatGPT to fix the errors but it doesn’t work at scale - it tells me it fixed all the errors but it fixed hardly any of them, even if I give it just 20 000 words or so.

A built in feature to correct OCR mistakes would be great but otherwise I’m just looking for any way to get these fixed that requires almost no work on my side.

1 Like
2

I do use Deepseek to correct incomplete texts with typo’s and so.
But even if it produces correct words in sé, these may not be the real words of the scanned text.

3

Please check out the new AI Polish feature which may help with this: NEW AI Polish feature

(the catch is that it’s currently limited to texts up to 7,000 words long)