Export your set from Quizlet as text, then paste that text into Paira. Quizlet documents its own export on the help page linked below, and whatever separator the export puts between term and definition, Paira detects it rather than asking you to declare it. There is no file conversion step in between.
Quizlet's own help page on exporting your sets
Why the format does not matter much
A vocabulary set is a small thing: a term, a definition, and sometimes an example. Every tool that holds one can emit it as lines of text, which is why moving between tools is a copy-and-paste problem rather than a migration problem.
Paira's importer tries five separators — tab, spaced hyphen, semicolon, comma, colon — and keeps whichever one splits the most lines into at least two non-empty fields. So an export separated by tabs and an export separated by commas both land correctly without you changing a setting.
How the importer reads a text file, separator by separator
What about CSV?
A CSV file is a text file with commas in it, so it imports the same way. The one thing to watch is definitions that themselves contain commas. If the first comma on a line is not the column break, that line will split in the wrong place.
The fix is to export with tabs if the tool offers it. A tab almost never occurs inside a word or a definition, so it removes the ambiguity that commas create. If you only have commas and your definitions contain them, a find-and-replace of the first comma on each line into a tab is faster than editing entries by hand afterwards.
If the export option is missing
People do run into this, and it is one of the more common complaints on the Quizlet subreddit. What is available on a given account and platform is Quizlet's decision, not ours, and it changes over time — which is exactly why the help page above is the authority on it rather than this post.
If you cannot export, you can still select the terms on the set page and copy them. Paste that into a plain text editor first. Copying from a web page often brings formatting with it, and a plain text editor strips it back to lines, which is all the importer wants.
What comes across, and what does not
Terms, definitions and example sentences come across, because they are text. Images, audio recordings and the study history you built up in the other tool do not — there is no format that carries them between unrelated products.
Losing the study history sounds worse than it is. Paira schedules reviews from what it has watched you do, so it starts every imported word at mastery zero and finds out where you actually stand within a session or two. A history imported from elsewhere would be a claim about your memory rather than a measurement of it.