So I started on the same work with the idea of extracting all relevant information from all cells (value ranges, values, formulas), exporting to TXT and then building a conversion layer. I didn’t get there but what was informative was the sparsity of information in a big spreadsheet. Most cells in the few models I had for testing have only a few that actually do meaningful stuff. Most are fluff. One of the yet untouched harder parts is the conversion of Excel-specific formulas. The Excels I had used mainly basic math and lookups (thousands of them), not many of the built-in Excel formulas. (Any tips on my approach are very welcome, VBA for Excel isn’t hard but quite hard to Google.)