Otherwise, it is a simple univariate function evaluator, not really a solver -- it doesn't handle multivariate equations, nor give numerical solutions to implicit equation systems F(x) = 0 (typically calculated using the Newton-Raphson method).
As far as I can tell, it doesn't seem to "solve" at all -- it merely evaluates the RHS, given a set of parameters.
It is definitely not a CAS or anything like that.
Also, you're in for a surprise if you search for Fibonacci. It gives a useless formula, and you cannot even evaluate the n-th number directly.
2) Finding all roots to a general nonconvex equation system is a NP-hard problem. It can be done by partitioning the feasible region into convex regions and doing local searches, but typically numerical methods will only return a single solution for a given initial guess. So it's fair that it cannot return all solutions.
3) The Fibonacci sequence is an iterative equation, so explicit iteration is typically required to evaluate a Fibonacci numbers at n. It's fair that a general root-finding algorithm cannot handle iteration.
I want to be fair to the authors, yet at the same time, want to provide gentle feedback that the technical depth of the product may not be sufficient to feasibly sustain a business (except maybe through ads? But even that will subside at some point).
Most engineers have large Excel files that do this sort of thing because it really isn't that hard to do (some use simple VBA code for Newton-Raphson -- it's Calculus 101 stuff), plus their Excel files integrate with their other custom calculations.