People have been proposing "pay as you consume" systems for decades but no one has ever figured out a sustainable business model. The overhead of accepting and distributing small payments appears to be too high, especially after dealing with fraud and chargebacks.
I don't think it's "no one has ever figured it out". It's more like, doing so would cost consumers less and companies would have less revenue. It's not an unsolvable problem. It's a problem no one - other than consumers - really wants to see solved. It's effectively a wallet. We can't solve the wallet problem? Wait! We have. Plenty of times.
Again, even at the solo site level it's not being done. Certainly, it could be.
Then start a business and do it. If your hypothesis is correct then there are thousands of websites that would love to accept payments from your service.
I think you'll find though that building a wallet is an extraordinary complex problem. The legal compliance issues alone make it tough to build a service that can operate in the USA. As soon as you build a service for transferring money, criminals will immediately try to use it for money laundering and other illegal payments. What's your plan to handle that? You can't just ignore it or most likely the government will shut you down. And those AML/KYC laws are unlikely to be loosened just to suit this use case.
Why? Why would I start a business where there is no market? We've gone over this. It's not a technical and/or solution issue. It's that the gate keepers - the ones who would but the product - don't want something that's good for consumers but bad for them.
https://www.nngroup.com/articles/the-case-for-micropayments/