- starts at -1,0 asymptotically to x axis
- maxes out at 0,1
- and back down again to mirror how it started.
Normal distributions keep tapering off well past the -1..1 endpoints, and I was looking for something simple that ended. And most importantly the formula for it!

Your conditions do not single out a particular function. For instance, $(x^2-1)^2, \frac{\cos(\pi x)+1)}2$ work. The first is a polynomial so that could be nice. The second is trigonometric.
If you want one that extends by 0 to a $C^\infty(\mathbb R)$ function, then the "bump functions" alluded to in the comment by D.B. would lead you to something proportional to $\exp(-1/(1-x^2))$.
Lets say $\mathcal F$ is the collection of functions $:[0,1]\to \mathbb R$ that satisfy the properties you laid out. Then I can see at least two things:
Here's some graphs on Desmos: