Thanks for the feedback.
I've actually thought about adding a solver that simply brute forces the solution by going through all the variations, but then I realized that there are 1606 octodecillion (4^100) possible combinations for a 10x10 maze, which I think is a "bit" much.
With a proper algorithm that solves the maze in a non-stupid way by discarding all variations that aren't solveable in the first place this is probably doable though. Might be a fun coding challenge when I have some spare time.