Stirling-PDF is the Swiss army knife for handling PDF’s. And it is free for small users and for selfhosted instances. A commercial subscription and enterprise service with support are also available. For most users it will replace commercial pc software completely.
Easiest way to run it, besides a free account on Stirling-PDF, is with a Docker instance on your local pc or on a intranet server. Behind a proxy server with basic authentication on the internet will do if you want to use it from other locations too.
OCR based on Tesseract works great although you need enough memory to handle large files. For a 100 MB file I need 4 GB total system memory available. With 2 GB my proxy server went down. Not tried myself but a Pi 4 with 4 GB should suffice. Needed storage size for files is limited because Stirling-PDF only create temporary files which are removed as soon as the files are downloaded.