Basic Workflow
The basic workflow for V‑Cloud API includes two steps.
Submit audio with desired transcription parameters via a form-field post to the
/transcribeendpoint. You will then receive arequestidstring for tracking purposes.Retrieve the resultant transcript using the returned
requestidand token.
Note
202 No result is currently available will return until the transcription process is complete.
Transcription Request
Use the following commands to submit a transcription request to the V‑Cloud API.
- Set a variable with a valid V‑Cloud token:
token=your-token-here- Set a variable specifying an audio file to transcribe:
file=/path/to/audioSample.wav- Submit audio for transcription with desired parameters:
response=$(curl -s -F "file=@$file;type=audio/x-wav" -F "token=$token" -F "model=eng1:callcenter" -F "emotion=true" -F "gender=true" -X POST https://vcloud.vocitec.com/transcribe) && echo $response && requestid=$(echo $response | cut -d\" -f4)
Fetch Transcription Results
- Command to get download link for results:
curl -sL "https://vcloud.vocitec.com/transcribe/result/$requestid?token=$token"
Sample Transcripts
JSON is the default output format for transcripts. For more information on different output formats, refer to output. The following sample is an unaltered JSON transcript.
{"started":"2020-04-03 14:28:59.457194","asr":"6.1.0-b1","donedate":"2020-04-03 14:29:22.975662","nchannels":1,"recvtz":["EDT",-14400],"streamtags":{"model":"eng1:callcenter"},"ended":"2020-04-03 14:29:22.976576","confidence":0.88,"utterances":[{"confidence":0.88,"end":1.34,"recvtz":["EDT",-14400],"start":0.05,"donedate":"2020-04-03 14:29:22.975662","recvdate":"2020-04-03 14:28:59.488266","events":[{"confidence":0.98,"end":0.41,"start":0.05,"word":"Hello"},{"confidence":0.84,"end":1.34,"start":0.5,"word":"earthlings."}],"metadata":{"source":"Hello_Earthlings.wav","model":"eng1:callcenter","uttid":0,"channel":0}}],"source":"Hello_Earthlings.wav","audiosecs":1.43,"model":"eng1:callcenter","recvdate":"2020-04-03 14:28:59.488266"}Optional — Make the JSON transcript easy to read by using the pipe operator to redirect transcript results to Python's json.tool module:
curl -sL "https://vcloud.vocitec.com/transcribe/result/$requestid?token=$token" | python -m json.tool
The result is a readable JSON transcript:
{
"started": "2020-04-03 14:28:59.457194",
"asr": "6.1.0-b1",
"donedate": "2020-04-03 14:29:22.975662",
"nchannels": 1,
"recvtz": [
"EDT",
-14400
],
"streamtags": {
"model": "eng1:callcenter"
},
"ended": "2020-04-03 14:29:22.976576",
"confidence": 0.88,
"utterances": [
{
"confidence": 0.88,
"end": 1.34,
"recvtz": [
"EDT",
-14400
],
"start": 0.05,
"donedate": "2020-04-03 14:29:22.975662",
"recvdate": "2020-04-03 14:28:59.488266",
"events": [
{
"confidence": 0.98,
"end": 0.41,
"start": 0.05,
"word": "Hello"
},
{
"confidence": 0.84,
"end": 1.34,
"start": 0.5,
"word": "earthlings."
}
],
"metadata": {
"source": "Hello_Earthlings.wav",
"model": "eng1:callcenter",
"uttid": 0,
"channel": 0
}
}
],
"source": "Hello_Earthlings.wav",
"audiosecs": 1.43,
"model": "eng1:callcenter",
"recvdate": "2020-04-03 14:28:59.488266"
}