Skip to main content
POST
Voice turn prediction check (is the speaker done, is it the same request, is it an interruption)

Authorizations

x-access-token
string
header
required

JWT token passed in x-access-token header

Body

application/json
phase
enum<string>
required
Available options:
listen,
drift,
end,
barge,
react,
warm
heard
string

The words heard so far (capped at 2000 characters). Required for every phase except warm.

earlier
string

The text an earlier held answer was started for (drift and end).

lastAssistant
string

What the twin said last, or is saying now for barge and react (capped at 600 characters).

specId
string
Pattern: ^[A-Za-z0-9_-]{1,64}$
requestArgs
object

Only the two fields that pick the conversation.

Response

The check's answer, or ok:false with an error.

ok
boolean
done
number

Probability (0–1) that the speaker is done (listen, end).

same
number

Probability (0–1) that a reply to earlier still answers heard (drift, end).

interrupt
number

Probability (0–1) that the user is taking the turn (barge).

react
object

How the speaking twin should react (react).

lookup
object
ms
number
error
enum<string>

react-off: voice reactions are not on for this session (react only).

Available options:
off,
react-off,
busy,
rate,
timeout,
error,
bad-input,
bad-answer,
bad-phase