That last part is exactly what my webcrawler does with asyncio plus threads: https://github.com/cocrawler/cocrawler

This is the part that sends work to a thread: https://github.com/cocrawler/cocrawler/blob/master/cocrawler...

I agree that this was confusing in the docs. Docs can be improved. It really helped that this isn't my first crawler written using cooperative multitasking.