As soon as I realized it was not a drop-in replacement

Forgive me for jumping to conclusions, but did you follow the SYNOPSIS ?

Specifically, in that Test::More::UTF8 was not written as a replacement but as an addendum . And thus, as the SYNOPSIS shows, you need to use both Test::More and Test::More::UTF8 .

I think it was the "is" and "isnt" functions that are used for the tests.

The error that it couldn't find "is" and "isnt", added to your phrasing of "drop-in replacement", is highly indicative that you did not include both modules, but tried to only use Test::More::UTF8; which is not sufficient for running tests.

Back when you were first asking about it, and that module was recommended for you, I was curious how it was doing the UTF-8 magic in a way that was different from the failed attempts, so I looked at its source: essentially, it just sets a few UTF-8 flags on the internal filehandles that the Test::More behind-the-scenes framework uses, and sets use utf8 for your script (unless you ask it not to). It's not defining any of the test functions (like is or isnt), and it relies on the underlying framework already existing so that it can modify flags (otherwise, it has no flags to modify).

On the filehandles: it apparently has its own handles instead of just using the plain-jane STDIN/STDOUT/STDERR, which is why your earlier attempts to set the binmode for the standard filehandles didn't help.