how to get stream of response by chunks without waiting when response done?

[AndrewBPC]

Like in https://chat.openai.com/
Or this package don't support steam?

[the-csaba]

Hi @AndrewBPC

You can set the stream parameter on the completions endpoint.

https://platform.openai.com/docs/api-reference/completions/create

When you specify stream=true in your request, you would then need to retrieve the response using the getResponse() method rather than toModel() or toArray().

You can then access the stream interface and loop over it, outputting it in chunks.

Unfortunately, we don't have any detailed documentation on that, but hopefully, this points you in the right direction.