I agree, i should not have hit the limits.. i certainly was not getting enough replies to warrant i was anywhere near hitting them. But low and behold i was rate capped.
So either this mod is wasting alot or google is lying about their caps...

Hey raz334, the mod uses the Gemma-24B model. According to Google’s pricing page, the daily cap is 14,400 requests (there’s no daily token cap, but there is a per-minute cap of 15,000 tokens).

What cooldown setting are you using? For example, if it’s set very low (like 2 seconds), the game could send around 30 requests per minute, which would use up the 14,400 daily cap in about 8 hours.

You can also check your usage in the Google AI Studio dashboard (where you got your API key), it’ll show exactly how many requests and tokens were used, which can help figure out what caused the cap.

it was indeed set low, as i was trying to experience the mod and wasn't really seeing a whole lot of new messages, leading me to believe it wasn't doing alot.
Does this frequently send messages perhaps to generate things that add to it for later use or something?
IE: Tell the AI something about the character or events, and then get a reply to save and use in another prompt later?

Even at the low setting i wasn't seeing alot of messages in game.

I have turned that up to it's full 30 seconds now, since i did check that dashboard and saw a huge spike during my game time... poof! there went the cap. It looked like context was where it hit its limit at.

https://photos.app.goo.gl/dSWharbULu52L4Vp7

my custom scenario with 10 pawns plus mechanoids, when i install them, sets off my local API limit. sometimes its the number of pawns and their interactions. smaller groups allow bigger API promps. largeer groups hit cap quicker with larger prompts. Just fyi on how to let play impact rimtalk api limits.