How do I create a workflow that solves captchas using OCR? Apple has a built in OCR feature when doing screenshots. There are other apps that have OCR features baked into Apple's screenshot feature.
I'm quite experienced with the OCR baked into KM, which is the Apple OCR feature encased in a KM action. I would have to see the captchas to be able to determine if it will work. Can you share an example (upload an image) of the type of captcha in question.
I don't want to get your hopes up. While Apple's OCR is super-fantastic, some captchas are pretty tough, even for humans. Moreover, Apple's OCR will never work with a single standalone digit, as far as my tests indicate. However some captchas are pretty easy to solve.
Will KM be able to solve these types of captchas?
No. This is way too difficult. Even when I look at them, it takes me 10 seconds to replicate what they are saying.
I've seen cases where Apple Text Recognition could solve things that humans couldn't, but that's usually when there was no maliciousness involved (eg, faint text.) Captchas are specifically designed to make it hard for OCR to read.