Merging CSV Files: Stacking Rows vs Joining by Column
"Merge CSV files" means two genuinely different operations depending on who's asking, and mixing them up is the single most common reason someone runs a merge and gets a result that's technically correct but not what they wanted.
Stacking: same structure, more rows
Stacking appends the rows of every file underneath one header row, the way you'd combine twelve monthly export files that all share the identical column layout into one file covering the whole year. If the same customer ID happens to appear in two of those files, stacking gives you two separate rows for that ID, not one combined row, because stacking never looks at the values inside the rows at all, only at matching up columns by header name.
Joining: different structures, one shared key
A join combines two files that have different columns but one column in common, a customer list and a separate orders list, joined on a shared customer ID, the same operation a database JOIN or a spreadsheet VLOOKUP/XLOOKUP performs. This needs the tool to know which column is the key and what to do when a key appears in one file but not the other, decisions a plain stacking merge never has to make.
What Kit-Bin's CSV Merge actually does
To be direct about it: CSV Merge only stacks. It appends the rows of every file you drop in under a single header row, matching columns by header name (so files don't need identical column order) and keeping every column seen in any file, filling a blank where a given file lacked one. It does not do a key-based join, and it won't silently produce join-shaped output that looks plausible but combined the wrong rows, if what you actually need is a join, this isn't the tool, and there currently isn't a Kit-Bin tool that does one.
If you need a join today
A spreadsheet's VLOOKUP or XLOOKUP formula, or a short script (Python's pandas merge(), or a SQL JOIN if the data's already in a database), is the practical route until a browser-based join tool exists. Stacking first with CSV Merge is still useful prep work if you first need to combine several files sharing one structure before joining that combined file against a second one.