Anthropic prompt caching / cache breakpoint

Hi,

Really enjoy BoltAI v2 so far, but one feature I miss from v1 is Anthropic prompt caching (cache breakpoints) for long system prompts or document analysis with Claude models.

Is support for this planned in v2? It was super helpful for keeping costs down and speeding up long-running chats. I’d love to see it added when possible.

Many Thanks,