[vishnukool]

I built basically this on top of my telegram setup last year. The thing that killed it was group chats, per-message LLM calls on a busy group eat your api budget alive, so I ended up putting a tiny local model in front as a first pass.

[mathigatti]

Yeah, in my implementation I pretty much ignore everything from groups except for explicit mentions to my telegram user, in that case that message is forwarded to claude to decide if sending a notification. That keeps the LLM calls pretty low.