Tokens that are used

[bikubay]

Is there is a estimate of how much more tokens would LangExtrac consume .e.g if we have an input of 1000 tokens, and the extracted concepts are 20 tokens. Given the multi pass that this has to do, should we be worried about exponential increases in token usage or is it a predictable cost