It’s a hassle in any code - whether that’d be three, babylon, ogl, Unreal, directly shaders, or any other option. Ofc you can just scale the UV values (here’s how to calculate them in the first place.), but with exception of very trivial shapes, that quickly becomes unmanageable, esp. if you do the scaling at buffers’ level and then start applying matrix transformations to said buffers, expecting UVs to react to that.
Hard to help without knowing the actual context of what you’re building, but if you’re stuck to coding UVs, triplanar mapping may be your best bet.