There is real literature suggesting biological signals carry information about compatibility. There is also a long history of consumer products making far stronger claims than the evidence supports, and the gap between those two is where this idea lives.
The open question. How strong would the evidence have to be before this is not pseudoscience?
What would kill it
The causal signal is likely weak relative to what a consumer product would need to claim, and privacy plus regulation raises the burden of proof before the first customer arrives.