> it always convert all the text to image on server side before serving.
It probably easy to recover to absolute precision with OpenAI-like models. I bet using AI could help a lot with recovering books in the future.
It probably easy to recover to absolute precision with OpenAI-like models. I bet using AI could help a lot with recovering books in the future.
On books that they do have exclusivity, though, people are doing exactly that, OCR the book and recreate the EPUB.
But it's not publically available since BW is very quick about patching their app once the crack is well known.