The reality is that even if you enter identical data, various spoke calculators will give you slightly different results.

The basic formula that most use will return the end of the spoke to the ERD you entered. However, there are some fudge factors not in the basic formula and the various programmers handle these differently with adjustments. There's the allowance for spoke elongation under tension (or not). There's the added length needed to allow for the effect of basket weaving the spokes at the last cross, and as flanges get large and you have crosses that put the spokes out away from radial, there's some added length not factored in the basic formula.

So for these reasons alone, the various programs will give you different answers. Then there are assumptions made a desireable margin of safety so you don't run out of thread, or so spokes don't end up short. Different builders have different biases here, and so do the programmers.

So my advice is to always use the same program so you have good sense whether it errs high or low and by how much. Once you're used to a program, and enter consistent data (measure yourself the same way), you'll add or subtract your own correction factor and hit the length dead on every time.

Since the OP is doing it for the first time, he should use identical data in 3-4 programs, and go with the weighted average. That should be close enough, and he'll then know how to adjust future calculations.