File object fields

Audio files of a call are returned in the files attribute of the call object. Each file is a JSON object with the following attributes:

Attribute	Type	Description
file_id	string	ID of the file, for example 01. It is unique within the call only. Pass it in the file_id parameter of Retrieve file for playback.
start_time	datetime	Date and time when the recording of this file started. ISO 8601 with time zone, for example 2024-06-05T01:35:55+03:00.
stop_time	datetime	Date and time when the recording of this file stopped.
file_size	integer	Size of the audio file in bytes.
watermark	string or null	SHA-1 hash of the audio file content. Use it to verify data integrity.
encrypt_key	string or null	Public encryption key. Set only when the audio is encrypted.
encrypt_tag	string or null	Encryption tag. Required to play back the encrypted audio. Set only when the audio is encrypted.
encrypt_fingerprint	string or null	Fingerprint of the encryption key. Required to play back the encrypted audio. Set only when the audio is encrypted.
speech_analytics	list of objects	Transcripts of the file. Empty when the file has no transcript. Returned only when the API user has the Call transcript view permission. See Transcript object.

Note

The location of the file on the storage is not exposed by the REST API. To retrieve the audio, use Retrieve file for playback.

Transcript object

Attribute	Type	Description
transcript_id	UUID	Unique ID of the transcript.
engine	string	Speech engine that produced the transcript, for example miarec_v2.
language	string or null	Language of the transcript.
ts_language	string or null	Full-text search language of the transcript.
status	string	Status of the transcript: UPLOADING (uploading audio), PROCESSING (in progress), RUNNING (waiting), COMPLETE, or FAILED.
transcript_plain	string	Plain text of the transcript. Empty until the status is COMPLETE.
transcript_json	string	Transcript with timing and speaker information, serialized as a JSON string. See Transcript JSON format.
created_at	datetime	Date and time when the transcript record was created.
updated_at	datetime	Date and time of the last modification of the transcript record.

Transcript JSON format

transcript_json is a JSON document encoded as a string. Parse it with a JSON parser to get a list of phrases. Each phrase has:

Attribute	Type	Description
side	integer	Speaker: 1 for the left channel (usually the agent), 2 for the right channel.
s	integer	Start time of the phrase, in milliseconds from the beginning of the file.
e	integer	End time of the phrase, in milliseconds.
c	number	Confidence of the recognition, when reported by the engine.
words	list of objects	Words of the phrase.

Each word has:

Attribute	Type	Description
w	string	The word, or a punctuation mark.
s	integer	Start time of the word, in milliseconds.
e	integer	End time of the word, in milliseconds.
sn	integer	Sentiment of the word, when sentiment analysis is enabled.
m	string	punc when the item is a punctuation mark. Absent for regular words.
sp	string	Spoken form of the word, when it differs from the written form (for example, a number).
ner	string	Named entity type, when named entity recognition is enabled.
ner_c	number	Confidence of the named entity recognition.

Older transcripts may also carry a per-word confidence c; new transcripts store confidence per phrase only.

Example (formatted for readability):

[
  {
    "side": 1,
    "s": 2880,
    "e": 6520,
    "c": 0.92,
    "words": [
      {"w": "Hi", "s": 2880, "e": 2920, "sn": 4},
      {"w": ",", "s": 2920, "e": 2920, "sn": 4, "m": "punc"},
      {"w": "I'm", "s": 3400, "e": 3520, "sn": 4},
      {"w": "calling", "s": 3520, "e": 3800, "sn": 4},
      {"w": "from", "s": 3800, "e": 3920, "sn": 4},
      {"w": "Delta", "s": 3920, "e": 4320, "sn": 4},
      {"w": "Air", "s": 4320, "e": 4600, "sn": 4},
      {"w": "Lines", "s": 4600, "e": 4920, "sn": 4},
      {"w": ".", "s": 4920, "e": 4920, "sn": 4, "m": "punc"}
    ]
  },
  {
    "side": 2,
    "s": 8480,
    "e": 9320,
    "c": 0.88,
    "words": [
      {"w": "Yes", "s": 8480, "e": 8520, "sn": 6},
      {"w": ",", "s": 8520, "e": 8520, "sn": 6, "m": "punc"},
      {"w": "this", "s": 8920, "e": 8960, "sn": 6},
      {"w": "is", "s": 8960, "e": 9080, "sn": 6},
      {"w": "she", "s": 9080, "e": 9320, "sn": 6},
      {"w": ".", "s": 9320, "e": 9320, "sn": 6, "m": "punc"}
    ]
  }
]