There are 2 types of elements available
1. GPT Streaming Element: This element takes a single message as an input of role being user. Setup instructions below

(i) Introduce the "GPT Streaming Element" in your page.
(ii) Add the "Generate Output GPT Streaming Element" action in any event (could be a click of a button, or a page load etc) in your Workflow tab.
(iii) Add your API key, select and pass the model in the action, and give the input to it, to which GPT would give you the output later on.
(iv) In the design tab, if you want to refer to the output being streamed, simply insert the dynamic content "GPT's Real-Time Output" in any input or text element.
(v) You are ready! Further enhance the experience by introducing the "A GPT Streaming Element Output Generation Completed" and set actions within it (could be to save the recently generated output).

2. Advanced GPT Streaming Element: This element has various options available. You can pass the JSON of messages to it as an input (with roles being system, user, assistant), the "Frequency Penalty", "Max Tokens", "N", "Presence Penalty", "Temperature", "Top_p". Setup instructions below

(i) Introduce the "Advanced GPT Streaming Element" in your page.
(ii) Add the "Generate Output Advanced GPT Streaming Element" action in any event (could be a click of a button, or a page load etc) in your Workflow tab.
(iii) Add your API key, select and pass the model in the action, and give the messages JSON to it, other optional parameters, to which GPT would give you the output later on.
(iv) In the design tab, if you want to refer to the output being streamed, simply insert the dynamic content "Advanced GPT's Real-Time Output" in any input or text element.
(v) You are ready! Further enhance the experience by introducing the "A Advanced GPT Streaming Element Output Generation Completed" and set actions within it (could be to save the recently generated output).