Authenticity verifications analyze the authenticity and liveliness of an audio to validate that the voice belongs to a human speaking live.

They can be used to detect a variety of spoofing or spam attempts, including the use of synthetic voices.

Combining authenticity verifications with message verifications can improve robustness against spoofing attempts.

Access: To manage a verification, you must have access to its space.

Isolation: A verification can only interact with resources inside its space.

Model: The analysis of the authenticity and liveliness of the voice is powered by a specific version of an authenticity verification model.

Scores: Confidence in positive results in different metrics is represented in the scores field.

 

Fields

---

idstringrequired
The UUID of the authenticity verification. Is unique.
Format: uuid
Example: 00000000-0000-0000-0000-000000000000

---

handlestringrequired
The handle of the authenticity verification. Is unique.
Example: verification-a

---

aliasstring | nullrequired
The alias of the authenticity verification.
Example: Verification A

---

statestringrequired
The state of the authenticity verification.
Format: VerificationState (string)
Possible values: [performing, performed, performedWithoutScore, error]
Example: performed

---

modelstringrequired
The model of the authenticity verification.
Example: default/v1

---

audio

object | null

required

The audio of the authenticity verification.

---

handlestringrequired

The handle of the audio. Is unique.

Example: audio-a

---

aliasstring | nullrequired

The alias of the audio.

Example: Audio A

---

statestringrequired

The state of the audio.

Format: AudioState (string)

Possible values: [avaitingFile, available, notAvailable]

Example: available

---

durationstring
The duration of the audio.
Example: PT3.500S

---

voiceDurationstring
The voice duration of the audio.
Format: duration
Example: PT3.500S

---

scorenumber | nulldeprecated
The score of the authenticity verification. This field is deprecated, use scores.synthesis instead. Only returned for the default and telephone models.
Format: float
Possible values: <= 1
Example: 0.85

---

subscores

object

deprecated

The subscores of the authenticity verification. This field is deprecated, use scores instead. Only returned for the default and telephone models.

---

synthesisnumber | nulldeprecated

The optional synthesis subscore of the authenticity verification. This field is deprecated, use scores.synthesis instead.

Format: float

Possible values: <= 1

Example: 0.85

---

replaynumber | nulldeprecated

The optional replay subscore of the authenticity verification. This field is deprecated.

Format: float

Possible values: <= 1

Example: 0.85

---

scores

object

required

The scores of the authenticity verification.

---

synthesisnumber | nullrequired

The synthesis score of the authenticity verification.

Format: float

Possible values: <= 1

Example: 0.85

---

chunksintegerrequired
The number of chunks of the authenticity verification.
Format: int32
Example: 10

---

createdAtstringrequired
The date and time when the authenticity verification was created.
Format: date-time
Example: 2023-01-01T00:00:00.000000Z

---

performedAtstringrequired
The date and time when the authenticity verification was performed.
Format: date-time
Example: 2023-01-01T00:00:00.000000Z

AuthenticityVerification

{
  "id": "00000000-0000-0000-0000-000000000000",
  "handle": "verification-a",
  "alias": "Verification A",
  "state": "performed",
  "model": "default/v1",
  "audio": {
    "handle": "audio-a",
    "alias": "Audio A",
    "state": "available"
  },
  "duration": "PT3.500S",
  "voiceDuration": "PT3.500S",
  "scores": {
    "synthesis": 0.85
  },
  "chunks": 10,
  "createdAt": "2023-01-01T00:00:00.000000Z",
  "performedAt": "2023-01-01T00:00:00.000000Z"
}