How to convert PDF to Excel without losing formatting?

If your reports are scanned PDFs, PaddleOCR’s PP-Structure is worth considering. It’s a free, open-source document-analysis toolkit that combines OCR with table-structure recognition. It runs locally, but it’s a developer tool rather than a simple desktop converter.

The workflow is to install PaddleOCR and the PP-Structure dependencies for your chosen version, download the required models, then run the table-recognition pipeline on your PDF pages. For a pipeline that accepts images, render the PDF pages as images first. Export the recognised tables using that version’s Excel export option.

Use one representative page as a test before processing everything. Check merged cells, multiline headings, and whether amounts landed in the correct columns. OCR can confuse characters such as 0 and O, so compare some figures and totals against the original.

It’s useful when you need to process many scans privately on your own computer. However, reconstructed table structure doesn’t guarantee identical formatting or that every exported value will already be an Excel number.