I've seen green roofs on bus stops and favelas in Rio, Brazil, where their roofs have terrible infrastructure. From what I understood, they basically partnered with a local university to try to reduce the heat island effect in favelas, as most of them had no greenery at all

They came up with a cheap solution involving a plastic liner, similar (perhaps the same) to those used in ponds; a geotextile layer of bidim, which is a lightweight material; and plants that don't need soil to grow, like those found growing on rocks or trunk of trees. Which means it is super light and still provide the same benefits as a green roof with soil in it

Eeeexcellent, I'm excited for you.

I've seen some science channels posting these over the last few years, making coatings that emit radiation in the "atmospheric window" I think it's called, but have not yet seen proper products on the market. I could be out of date though. Still, I'd question durability, but I suppose it'd be at least no worse than normal materials, right?

Unnecessarily reductive. It's both the ceilings and walls. There is no single solution. We also have no idea of they have an attic or vaulted ceilings. An attic full of hot air is a heat bank that will radiate into the living space all night. All insulation will do is slow the transfer. Really good attic floor insulation will help retain heat in the attic, so you still want to improve attic ventilation. But, if you reduce the amount of heat going into the attic in the first place, you reduce how much insulation and ventilation you need to reach a certain cost/benefit ratio. It's usually more feasible to do half a job addressing each than it is to make just one variable perfect. So, when it comes to replacing a roof covering that needs replacing anyway, doesn't it make sense to do that as best as feasible?

Tour a home with an infrared thermometer camera in summer. I promise you, the ceiling is warm, too. You just don't hang out 3 meters up.