I don't think we want a new arg, but maybe just an option, that we can set for markdown reports for instance.
We set a token amount in this threshold, if it has been overtaken during the last minute we wait.
For instance if 50% of tokens are taken by the last query, wait so the next query has a sufficient window.
I don't think we want a new arg, but maybe just an option, that we can set for markdown reports for instance.
We set a token amount in this threshold, if it has been overtaken during the last minute we wait. For instance if 50% of tokens are taken by the last query, wait so the next query has a sufficient window.