1. when you set startSpeakingPlan options, how do waitSeconds, smartEndpointingEnabled, and transcriptionEndpointingPlan interact? e.g. If I set smartEndpointingEnabled = true but don't set the other 2 options, are they still active at their default values? 2. Is responseDelaySeconds deprecated? Thanks!

Optional
This is the plan for when the assistant should start talking.

You should configure this if you’re running into these issues:

The assistant is too slow to start talking after the customer is done speaking.
The assistant is too fast to start talking after the customer is done speaking.
The assistant is so fast that it’s actually interrupting the customer.

Hide 3 properties
waitSeconds
double
Optional
This is how long assistant waits before speaking. Defaults to 0.4.

This is the minimum it will wait but if there is latency is the pipeline, this minimum will be exceeded. This is really a stopgap in case the pipeline is moving too fast.

Example:

If model generates tokens and voice generates bytes within 100ms, the pipeline still waits 300ms before outputting speech.
Usage:

If the customer is taking long pauses, set this to a higher value.
If the assistant is accidentally jumping in too much, set this to a higher value.
@default 0.4

smartEndpointingEnabled
boolean
Optional
Defaults to false
This determines if a customer speech is considered done (endpointing) using the VAP model on customer’s speech. This is good for middle-of-thought detection.

Once an endpoint is triggered, the request is sent to assistant.model.

Default false since experimental.

@default false

transcriptionEndpointingPlan
object
Optional
This determines how a customer speech is considered done (endpointing) using the transcription of customer’s speech.

Once an endpoint is triggered, the request is sent to assistant.model.

@Ethan Tan 1. When you set the

startSpeakingPlan

options,

waitSeconds

,

smartEndpointingEnabled

, and

transcriptionEndpointingPlan

work together to determine when the assistant should start speaking. If you enable

smartEndpointingEnabled

by setting it to

true

but do not explicitly set

waitSeconds

or

transcriptionEndpointingPlan

, they will still be active at their default values. The default for

waitSeconds

is 0.4 seconds, and

smartEndpointingEnabled

is false by default, which means if you set it to true, you are overriding the default behavior. The

transcriptionEndpointingPlan

has its own defaults if not set. 2. As per the information provided,

responseDelaySeconds

is not mentioned as deprecated. It is a separate configuration that specifies the minimum number of seconds to wait after user speech before the assistant starts speaking, with a default value of 0.4 seconds. If there have been updates to the API that deprecate this option, it would be best to refer to the latest documentation or release notes for confirmation. If you need further assistance on this matter, please ask support for more info on this.

1. Interaction of

waitSeconds

,

smartEndpointingEnabled

, and `transcriptionEndpointingPlan`: -

waitSeconds

provides a basic delay before the assistant starts talking, acting as a minimum wait time. -

smartEndpointingEnabled

takes priority if enabled. When

smartEndpointingEnabled

is

true

, it overrides

transcriptionEndpointingPlan

entirely, meaning the assistant uses a smart model for timing and ignores other endpointing settings. -

transcriptionEndpointingPlan

only applies when

smartEndpointingEnabled

is

false

. It adjusts timing based on transcription patterns like punctuation or numbers. 2. If you set

smartEndpointingEnabled = true

but don’t set the other two: - Yes,

waitSeconds

and

transcriptionEndpointingPlan

are ignored entirely when

smartEndpointingEnabled

is

true

. 3. Is

responseDelaySeconds

deprecated? - Yes,

responseDelaySeconds

is essentially replaced by

waitSeconds

in the

startSpeakingPlan

configuration. Use

waitSeconds

instead for setting the initial delay before the assistant speaks.

thanks @Shubham Bajaj