Why the columns are padded, not just piped#
A Markdown table renders identically whether the source is padded neatly or crammed with no extra spaces — |a|b| and | a | b | produce the same table in any renderer. The padding exists entirely for the person reading the raw Markdown source before it renders: a column-aligned table in a .md file is legible on its own, in a diff, or in a terminal, while an unpadded one is a wall of pipes. This tool pads every column to the width of its longest cell, the same convention most hand-formatted Markdown tables follow.
How the delimiter is detected#
Selecting cells in Excel or Google Sheets and copying them puts tab-separated data on the clipboard — pasting that directly here is detected automatically, since the first line will contain a tab character. Plain comma-separated text, with no tabs, is parsed as CSV instead. There is no manual delimiter switch to set; the tool reads the first line and picks accordingly.
What the colons in the separator row actually do#
The line of dashes under the header row is what makes a Markdown table a table rather than plain text — GitHub-flavored Markdown additionally reads a colon on either end of that dash run as an alignment hint: :--- for left, :---: for center, ---: for right. Plain --- with no colon defaults to left-aligned in every renderer that supports table alignment at all. This tool writes the correct colon placement automatically based on the alignment chosen per column.