I have the exact same idea for Symantic embedding based content aggregation (RSS reader + assistant)

What was your stack for processing all RSS pages on the internet ?