Choose needed models from repositories

Fast tessdata

Standard tessdata

Best tessdata

or download any other %.traineddata for tesseract and rename it according pattern {language_code}.traineddata, examples can be found in repos above

Create Zip archive to import

Structure should be following

• root (zip archive)
  • standard (package)
    • tessdata (package)
      • {...}.traineddata (your models, any count)
  • best (package)
    • tessdata (package)
      • {...}.traineddata (your models, any count)
  • fast (package)
    • tessdata (package)
      • {...}.traineddata (your models, any count)

archive must contain at least one of (standard, best, fast) packages with models

Import created ZIP archive

this can be done in language selector in OCR tool

Done! You can start recognition