Ok, senpaitachi. I think I managed to come up with a scent version of a microphone and speaker. I call them the "smeller" and "disperser".

If I get the anatomy right, there’s a finite number of scent receptors (this could be incorrect terminology, but the meaning is probably right). Around 500, or perhaps more. 500 for the sake of argument.
We could look at what reactions occur to detect the individual chemical, or range of chemicals, that each scent receptor detects. Then we could copy the mechanism involved, or we could simply use already established tests for particular chemicals or ranges of chemicals.
This would take 500 different artificial receptors or so.
When we had the information from these 500 different receptors about how much of their chemical or range of chemicals was in a sample of air, we could store this information.
We could then have a collection of 500 scent dispersers that would each dispense one chemical that generates one of the 500 different basic scents.
We could use the information from the smelling machine to disperse an appropriate amount of each of the 500 scent chemicals (from none to the maximum amount that the air can hold).
This should be a complete “scent microphone and speaker”.