The tagline is too close to "Fastest Fourier Transform in the West" or FFTW [1].

The name "WebFFT" is fine, just do away with the "Fastest..." bit. It looks like the author is trying to piggy-back on the FFTW brand.

[1] https://en.wikipedia.org/wiki/FFTW

Using a list of heap allocated objects for a 2d FFT is an odd choice.

It's also worth checking out the project that the creator uses it in: https://github.com/IQEngine/IQEngine

How does it compare with https://github.com/webfft/webfft.github.io?

this should be faster https://github.com/webfft/webfft.github.io technically because it is using hardware for fft whereas current library is doing them on cpu but there is additional overhead with hw so we would actually have to test it.

Looks like this one can do 2D transforms.

The title is an obvious reference to fftw.

I'll have to try it out, I've used a transpiled kissfft until now.