What's a term for pure text data (non-image non-video non-audio)?

I would normally call this "plain text".

In computing, plain text is the contents of an ordinary sequential file readable as textual material without much processing, usually opposed to formatted text. - wikipedia

The data type which includes image, video and audio is named as "binary data". So, if you are storing data that you want to represent as text, such as a blog post or product description you may use "non-binary string data".

If you want to store data that are merely bytes such as data that represents image, then you use a binary string data type.

I'd say "text data". If you want to be a lit more technical, I'd say "text and numeric data". I'd consider XML or other marked-up text to still be "text". When we need to distinguish between HTML, XML, RTF, etc and text without markup, we normally say something like "plain text" versus "marked-up text" or "tagged text".

"Text data" is a common term that anyone in the business would understand -- possibly with some ambiguity about whether it includes XML or a Word doc or that sort of thing, but they would surely understand that it does NOT include images, audio file, video, binary code, etc. A non-professional would have a reasonable chance of figuring out what you mean even if you didn't explain further.

So sure, you could invent some phrase that would be more technically precise, but no one would know what you meant unless you explained it. And if you're going to explain it, you might as well use a simple, common term and explain that. Save inventing new words for when you've really invented a new concept.