

You could always self host something like lm studio and download some models of interest and adjust to the highest amount of input. Of course you will be limited by the hardware as well as the model. As an example, I don’t think LM Studio imposes a limit, it depends on the model.
My apologies if I misunderstood any of your question.
LM Studio is free
Have you tried notebookLM from Google yet? Provide it s couple documents or web pages and then have it generate an audio analysis. You will get a podcast style report with two voices. They have some pretty advanced production. You can even hear them breathing.