I built Umee PDF Pro because I needed to extract return tables from fund factsheets . I kept running into the same wall: every tool either uploaded the file to a cloud server (not acceptable for client data) or couldn't detect the table structure reliably enough to be useful. So I built it myself. Started as a side project to solve my own workflow problem in finance. The table extraction engine took the longest. Now I'm opening it up. It's live at umee.online as a web app.