We get the following a status error alert on some pages, with the quest response returning:
{
"status": false,
"error": "This model's maximum context length is 16385 tokens. However, your messages resulted in 68396 tokens. Please reduce the length of the messages."
}
First, the error should also be shown as flash message, and not just an alert without a message, see #2.
Second: How can this be resolved from an integrators perspective?
We probably could strip a few more parts of a page from the rendered markup before markdown conversion.
We get the following a status error alert on some pages, with the quest response returning:
{ "status": false, "error": "This model's maximum context length is 16385 tokens. However, your messages resulted in 68396 tokens. Please reduce the length of the messages." }First, the error should also be shown as flash message, and not just an alert without a message, see #2.
Second: How can this be resolved from an integrators perspective?
We probably could strip a few more parts of a page from the rendered markup before markdown conversion.