I'm also getting error messages for the Chrome plug in.

For the past hour I've had intermittent problems with nbib files that were badly created at PubMed. A few moments later then requesting the same file it is fine. I opened two bad files in a text editor and at the beginning there were several gibberish characters (different gibberish strings in each). I sent a tech support report to NLM.

edited to fix typos and to add a word

I had lots of "�" characters at the beginning of the files along with what appeared to be old-style block-graphics characters and card symbols. There were numerous extraneous line feed characters (at least, that is what my BBEdit program displayed).

There were many of these inappropriate characters throughout the file.

edit: It was clear that these "funny characters" had replaced normal character strings throughout the file. This was clearly an NLM problem that has nothing to do with Zotero.

Here are the first three visible lines of the flawed file that I recovered:
���□������□□
��□▤ ▫◘ ◩ 3315▧ ▤


O◪N▧▧�◩LM

Here are the first three lines of the same nbib download that I was able to import to Zotero:

PMID- 34333159
OWN - NLM
STAT- Publisher

As I paste the flawed lines from my text editor to this forum message the "funny" characters aren't always displayed the same way (the card symbols in my editor became white on black question marks)-- but you get the idea.