Hi,
Do you mean that you set the max_tokens parameter? If yes, that’s not the way to do it ?? This max_tokens parameter is the maximum number of tokens the AI model is authorized to generate (actually, the number of tokens you send through the prompt is also counted). So yeah, as the discussion goes, it might be way above that.
That said, to remember the discussion, the entire discussion needs to be sent (same with ChatGPT actually, it does that too). The AI itself just takes one input (the prompt) and gives one output. So every time the user sends a new question in the same discussion, the previous questions/answers are also sent, and are counted as part of this max_tokens parameter.
I think you needed to know what that parameter really is, as it’s an important concept.
Now, for your limitations, it’s a bit trickier, and you can’t do it this way. The way to do it is explained here: https://meowapps.com/ai-engine/tutorial/#usage-control
I’ll make more examples based on the use cases of my users. Don’t hesitate to share yours, and why you would like to have a different page for the chatbot. Is that page behind a paywall? Even though there is a paywall, what happens after, when the tokens are consumed again? Let me know your use cases, I’ll try to provide an example ??