Comment by late_night_fix

I try to batch requests and reuse responses wherever possible.LLM API calls are awesome,but if you don't watch your usage,cost sneaks up fast.