NP-complete problems are decision problems and belong to NP (and every problem in NP can be reduced in polynomial time to them, but these details I guess you already saw online).
NP-hard are problems to which any problem in NP can be reduced, but not necessarily belong to NP or are decision problems.
Obviously, every NP-complete problem is also NP-hard (by definition of NP-hard). The opposite is not true, there are problems that are NP-hard but do not belong to NP.
For example, finding count of all solutions to a SAT instance (#SAT) is NP-hard but does not belong to NP-complete class, at least because it is not a decision problem and hence does not belong to NP.
On the other hand, SAT, the problem of deciding if count of satisfying solutions is greater than zero, belongs to NP and every problem in NP can be reduced to it, hence it is NP-complete.
Note, every problem in NP can be reduced to (#SAT) (because SAT can be reduced to #SAT, just find a count and output true if it is non-zero). It is "hard" at least as SAT; this is the intuition behind the name NP-hard.
I would also like to point to an excellent and detailed answer covering more details.