Constructing and Manipulating Raw Strings

[Nordlöw]

What is the best way to extract the individual bytes from hexadecimal
string (output from md5) and using them as input into a call to
unibyte-string()?

I want to use this raw string as a key in a hash table for file types
and then I don't want to waste space (size halves if we go from a
hexadecimal string to a raw byte string).

Does read() (or some of its variants) support reading raw unibyte
strings (that may contain zero bytes).

Thanks in advance,
Per Nordlöw