XLSX Repair

An .xlsx Is a ZIP, and This Checks Its Parts

Since Excel 2007, an .xlsx or .xlsm file is an Office Open XML package: a ZIP archive holding one XML file per piece of the workbook. This page opens that archive in your browser and reports which pieces are present, which required ones are missing, and whether the workbook as a whole still parses.

Two more parts are worth knowing about even though the checker does not require them: xl/sharedStrings.xml holds the text values the sheets point at, and xl/styles.xml holds number formats, fonts and fills. Both are listed in the report only indirectly - if either is damaged, the parse step is where it shows up.

Structure Intact, Content Broken - and the Other Way Round

The two failures look identical in Excel and are completely different here. When the archive opens and every required part is present but the parser still cannot build the workbook, the status is repairable: the ZIP is fine and the XML inside one of the parts is not. When a required part or every worksheet part is gone, the status is limited, and no amount of re-parsing invents the missing bytes. When the ZIP itself will not open, the status is unsupported - the archive is truncated, the file is encrypted, or it was never an .xlsx.

Extension Says .xlsx, Contents Say Otherwise

A large share of "corrupt XLSX" files are not corrupt at all. Exports from web apps, CRMs and reporting tools are frequently CSV or HTML tables handed out with an .xlsx name, which is exactly what produces Excel's "file format or file extension is not valid" warning. The give-away is the unsupported status here plus readable text at the start of the file when you open it in a text editor. Rename the copy to .csv or .html and convert it properly with CSV to Excel or HTML to Excel.

What the Recovery Export Produces

When the workbook parses, the export walks every sheet, reads the cell values, and writes them into a new workbook named yourfile-recovered.xlsx. It is a values-only rebuild: formulas land as their stored results, dates as serial numbers, and styles, charts, images, pivot tables and macros are not written at all. Compare the row and column counts against what you remember of the original before you rely on it - a sheet whose XML was truncated can parse successfully and still stop halfway down.

Nothing here is uploaded: the ZIP is opened, inspected and rebuilt inside the page. That also means it works on a workbook you are not allowed to send to an online repair service.