Some examples of this might be Standard American English (though this may still be tied to geography) or, more likely, Received Pronunciation. The speaker's language doesn't have to be English, of course.
Basically, the speaker's accent isn't discernible or particular to some region. It may be that the speaker has a unique way of talking or a blend of accents such that a listener can't place the speaker's origins. I don't think "accent-less" would be valid, since technically there's no such thing as having no accent. I'm also trying to describe the (artificial? synthetic?) accent of some text-to-speech programs.1
1: Can a TTS program truly have an accent in the first place? (I know you can assign an American or Australian "accent" in some cases, but I'm referring to the crude, "robotic"-sounding speech typical of early TTS.)
Best Answer
FrustratedWithFormsD's suggested indiscernible implies you're not aware of any accent at all. Which is increasingly common today. Just as genetically we're all becoming coffee-coloured people in this modern world of global communications and travel, so differences in regional accents tend to be "ironed out".
But OP wants a word to describe speech which is clearly recognisable as having some kind of accent - just not a "placeable" one. Perhaps because someone's speech combines two or more relatively strong accents in a way that makes it difficult for others to identify the components.
I'd call that a nondescript accent, as apparently would many others in that link.