with the conversation history to the

[hasinam2206]

Js helper library, starts a conversation if none exists, listens for user input, and passes that input along endpoint respondfor processing. On line 6, the application checks if a called cookie convoexists. If it doesn't exist, or if it exists but is empty, you can assume that the conversation hasn't started yet.

So you'll send a first welcome message using the 11-digit phone number format philippines twiml verb. Then the method twiml.gatheris used to capture the data entered by the user. The parameters for gather are as follows: speechtimeout: 'auto': automatically determines when the user has stopped speaking. It can be set to a positive integer, but "auto" is better for this use case.



Speechmodel: "experimental_conversations": : uses a speech recognition model optimized for conversational use cases input: 'speech': sets the input type to "speech" and ignores keystrokes (dtmf) action: '/respond': transmits the user's voice input along with the conversation history to the endpoint/respond now, you need to make sure to create the convo cookie so that the function /respondhas a place to store the conversation history that will be passed between the openai api and our polly neural voices.