If you haven't noticed already, Claude has officially cut off Open Claude from using their subscription service, so you can no longer use the Max plan or the Pro plan, which is $100 a month for basically unlimited usage. And you have to go into API keys now, which I've basically already worked out I'd rack up about $1,000 worth of uh usage of tokens. So, instead of doing that, you need to set in what's called an orchestrator layer. This orchestrator layer enables you to basically choose which models to use based off the requirement of the request. So, first up, head into Claude, get it to put together a strategy to add to my server. Here's the basic setup I'm going to use uh Opus, Sonnet, GPT-4 Mini, and then also it's not put in there, I'm going to use Codex as well, which is 5.4. Transferred that over into my server, the MD file, and let it run its magic via Claude code on the server. Then once it's configured to make sure it's working correctly, it's got 5.4 installed. We've got Sonnet and Opus. Now, whenever I ask via my Telegram for it to do something, it will check which is the best model to use according to the instructions that I've given it, and hopefully that keeps me within a sane amount of money that I'm spending on my AI subscriptions. Hope that helped.