Ocr API fails from get request

Im having issues with this image. 1. when i use get request im getting
an error suggesting urlencode image url. Though there is no
documentation on how to do that on your site and when i do it with
urllib.quote - it returns nothing at all. 2. when i upload image or use
url on your site it does not return results or returns incorect one. So
can your service correctly recognize an image like this? How do I do
this in my code?

Have you tried the new OCRengine=2 option yet? I tested it with your captcha and it works “ok” (not perfect, but better than engine1).

See Improved alphanumeric OCR and special character OCR