Will this produce a uniform distribution?

((integer)("0x" + llGetSubString(llMD5String(inString,0), 0, 7))) % p, for p a prime.

In theory it should if your input set is large enough (i.e. if your input strings are only one character you'll probably only get about 128 results, and if p>128...). Keep in mind though that the modulo operator will produce a slight bias itself since 2^32 is not divisible by p (except for the trivial case where p=2). If p is large enough, this could become significant. For example, if p is somewhere around (2/3)2^32 then the first p/2 values are twice as likely as the last p/2 values. I've seen it recommended that you actually use division rather than modulo for pseudo-random number generation for this reason.

Excellent! Thank you.