I’m trying to learn how to make use of ollama and local LLMs to do some co-processing for coding/doc review/etc so I’m not burning through tokens. But now I have to read about how this open-claw changes with Claude and having to pay-as-you-go tokens will affect ollama usage too
My favorite custom command is /fresh. It’s write a hand off doc, pushes to GitHub. Then I kill the session and start over. I know the context window is massive but I only want to pass the context I need per turn. I have found this has massively cut my token usage and has made help enforce discipline over what I tackle per session
This is helpful. I’ll try these out.
This is extremely helpful.
yay!
Merci!
That’s useful will try that.
Thank you for sharing. These Claude tips arrived on time 👌
Can these steps be applied to cowork?
Love this! Thank You so much!
I’m trying to learn how to make use of ollama and local LLMs to do some co-processing for coding/doc review/etc so I’m not burning through tokens. But now I have to read about how this open-claw changes with Claude and having to pay-as-you-go tokens will affect ollama usage too
So helpful, love it
Another point is to off unused connection! If you don’t use connection to jira - turn it off, for example. It will free space in context.
I think your tips are going to save me some money!!
Thank you ❤️
My favorite custom command is /fresh. It’s write a hand off doc, pushes to GitHub. Then I kill the session and start over. I know the context window is massive but I only want to pass the context I need per turn. I have found this has massively cut my token usage and has made help enforce discipline over what I tackle per session