There isn't strong evidence, and you can't just eyeball studies and assume you understand them with no training.
I have cystic fibrosis, and I have tried to do things like this before by looking at studies on unverified treatments. It doesn't work. You just end up with snake oil. I have a masters in biomedical engineering, and I couldn't do it.
Interpreting studies is hard. Interpreting meta analysis accurately is super hard. Tons of studies are fake, or have bias, or have poor methodology. That website is total bullshit and is just throwing all the studies in to their statistical soup and mixing them together to come up with nonsensical results.
"There is evidence of a negative publication bias, and the probability that an ineffective treatment generated results as positive as the 65 studies is estimated to be 1 in 403 billion."
Lol, wtf?
I'm no medical anything but I do have extensive academic and statistical training. There is enough in these studies for someone to take Ivermectin as an "eh, why not?" gambit. There is evidence it is helpful, although the evidence isn't conclusive (or even especially strong). Some doctors are already prescribing it, quite reasonably.
It certainly could turn out to be nothing. But on the other hand, why not get a doctors prescription until a couple of high quality studies have decided the issue? As you point out, the only real problem seems to be people taking it in unsafe dosages or through unreliable channels like a vet.
[0] Which is broadly consistent with the studies I've seen over at https://ivmmeta.com/ - not sure by how much, but it does seem to assist.