Hi,
I'm currently working on my project: https://github.com/Maciek-roboblog/Claude-Code-Usage-Monitor, and I've been exploring how to handle token limits (max_tokens) in a more generic way — especially when it comes to tracking total token usage across the entire flow: input, output, read/write cache, etc., and converting all of that into a unified measure like total "generated" or consumed tokens.
I was wondering if you’ve come across this topic or have any insights, data, or experience dealing with this kind of token accounting? Would be great to exchange ideas — or even collaborate if it makes sense.
Hi,
I'm currently working on my project: https://github.com/Maciek-roboblog/Claude-Code-Usage-Monitor, and I've been exploring how to handle token limits (max_tokens) in a more generic way — especially when it comes to tracking total token usage across the entire flow: input, output, read/write cache, etc., and converting all of that into a unified measure like total "generated" or consumed tokens.
I was wondering if you’ve come across this topic or have any insights, data, or experience dealing with this kind of token accounting? Would be great to exchange ideas — or even collaborate if it makes sense.