You might want to check out Docspell - it’s lighter than paperless-ngx but still handles PDF indexing and searching realy well, plus it can do basic OCR on those image-based PDFs without much setup.