Detect Scam Call event
The events sent to the client on the Detect Scam Call API.
- DetectScamCallResult
- TranscriptEvent
- SessionMetadata
- ErrorEvent
An event emitted when processing the audio stream has yielded a result.
The overall scam risk of the call, based on the callOriginatorRisk, scamDialog and syntheticVoice factors.
CallScamRisk (string)
Possible values: [NOT_SCAM
, MEDIUM_SCAM_RISK
, HIGH_SCAM_RISK
]
MEDIUM_SCAM_RISK
The risk associated with the call originator, regardless of the content of the call.
CallOriginatorRisk (string)
Possible values: [UNKNOWN
, LOW
, MEDIUM
, HIGH
]
MEDIUM
scamDialog
object
required
The scam risk based on the speech content in the audio stream.
ScamDialogRisk (string)
Possible values: [SCAM
, NEUTRAL
]
SCAM
The confidence level of the scamDialogRisk field.
ScamDialogConfidence (string)
Possible values: [LOW
, HIGH
]
HIGH
syntheticVoice
object
required
Whether the caller in the audio stream has been detected as a synthetic voice.
SyntheticVoiceDetected (string)
Possible values: [YES
, NO
]
YES
The score associated with the syntheticVoiceDetected field. 0 means total confidence in the voice being synthetic, 1 in being non-synthetic.
double
0.1
Type discriminator.
result
result
An event emitted when a chunk of the audio stream has been transcribed.
Whether a line of the transcript was spoken by the user or the non-user party.
TranscriptSpeaker (string)
Possible values: [USER
, SUBJECT
]
USER
The transcript line corresponding to a section of the audio stream.
hello good afternoon
Type discriminator.
transcript
transcript
The event emitted when the session has been established, as a response to the initial metadata frame sent by the client.
The unique identifier of the session.
Type discriminator.
sessionMetadata
sessionMetadata
An event emitted when an error has occurred.
The error message.
Type discriminator.
error
error
{
"callScamRisk": "MEDIUM_SCAM_RISK",
"callOriginatorRisk": "MEDIUM",
"scamDialog": {
"scamDialogRisk": "SCAM",
"confidence": "HIGH"
},
"syntheticVoice": {
"syntheticVoiceDetected": "YES",
"score": 0.1
},
"type": "result"
}