Can I paste directly from Excel or Google Sheets?
Yes. Copied cells paste as tab-separated text, and Auto detection picks that up, so you get one JSON object per row with your header row as the keys.
How are quoted fields handled?
Fully. Fields in double quotes can contain commas, line breaks and doubled quotes ("") for a literal quote mark, following the standard CSV format (RFC 4180).
What does “Detect numbers” do?
It turns values that look like numbers into JSON numbers and true/false into booleans, but only when nothing changes: 007, 1.50, 1e3 and numbers too long to store exactly, like 12345678901234567890, stay strings. Turn it off to keep every value as a string.
What if a row has more or fewer cells than the header?
Missing cells become empty strings and extra cells are kept under the names column4, column5 and so on, so nothing is lost. The status line says how many rows didn't match the header.
What if two columns have the same name?
Each name has to be unique in a JSON object, so repeats are numbered: two columns called a become a and a_2, and a third becomes a_3. The status line tells you when that happened, so no column is lost.
What if there is no header row?
Turn off “First row is header” and each row becomes an array of values instead of an object.