Add-on

VoiceBase Transcription w Custom Vocab

Automatic Transcription with Custom Vocabulary to Improve Business Processes

Increase the value of your Twilio audio recordings with the world's most accurate and incredibly fast, speech recognition solution. Automatically detect and discover what's said, using keyword and topic extraction to garner a greater knowledge about your business. Twilio developers can now automate important business processes based on what customers are saying to save resources, improve customer experience and better the agent training process. Transcription and keyword analytics are returned from VoiceBase in JSON data within 2 hours, for faster turnaround.

  • Accurate Transcription

    Recordings sent from Twilio are quickly processed in parallel by VoiceBase, using a proprietary speech engine to deliver industry-leading accuracy. VoiceBase returns a JSON response with a time aligned transcript, TXT and SRT formats and with speaker identification if dual channel is utilized.

  • Keyword and Topic Discovery

    VoiceBase automatically extracts the keywords and topics that are most important to your content. Discovered keywords and topics are delivered under the keywords object in the JSON response. Keywords include the start "s" and end "e" times and position "p" of each keyword in the transcript.

  • Custom Vocabulary

    The VoiceBase speech recognition may be augmented and optimized with custom vocabulary terms. Terms recognition may also be enhanced by adding a weight (0 - 5) or Sounds Like terms. Enter each term with the following structure: "term;sounds like term;weight Ex. "Twilio;twilleeeooh;3"

  • Keyword Spotting

    You can define keywords (or key phrases), which are flagged when they are spotted in the recording. Each keyword or phrase is input as a string separated by a comma in the Keyword Spotting field of the Configure section of the Add-on.

Sample annotation responses

Transcription Response

Transcripts are sent back with the format requested txt for plain text transcripts and srt for closed caption transcripts. Each word and punctuation is individually identified with end "e", start "s", confidence "c" and position "p".

{
"media": {
  "transcripts": {
      "text": "This is a test recording. ", 
     "words": [ { "w": "This", "e": 2210, "s": 1760, "c": 0.7, "p": 0 },
                { "w": "is", "e": 2290, "s": 2210, "c": 1, "p": 1 },
                { "w": "a", "e": 2760, "s": 2290, "c": 0.759, "p": 2},
                {"w": "test", "e": 3560, "s": 2760, "c": 0.533, "p": 3},
                {"w": "recording", "e": 3639, "s": 3560, "c": 0.501, "p": 4} ],
       "srt": "1\n00:00:01,76 --> 00:00:07,97\nThis is a test recording."
   }
 }
}

Topics Response

Topics discovered by semantic indexing topics are listed with the keywords that fall under each topic.

{
  "media": {
    "topics": {
      "topics": [
        {
          "name" : "Law of obligations",
          "score" : 0,
          "keywords" : [ 
            { 
              "name": "Payment",
              "internalName": [ "payment" ],
              "t": { "unknown": [ 116.71, 185.72 ] },
              "score": 1
            } 
          ],
          "type" : "category",
          "similarCategories" : [   ],
          "subcategories" : [  ],
          "speakers" : [ "unknown" ]
        } 
      ]
    }
  }
}

Keywords Response

Keywords from semantic indexing are provided with the time positions, broken down by speaker followed by a list of spotted key words.

{
  "media": {
    "keywords": {
      "words" : [ 
        { 
          "name": "air conditioning service",
          "t": { "unknown": [ 197.69, 200.9 ] },
          "relevance": 0.982013790038
        } 
      ],
      "groups": [
        {
          "keywords": [
            {
              "name": "data science",
              "t": { "unknown": [ 4.62 ] }
            },
            {         
              "name": "data mining",
              "t": { "unknown": [ 6.68 ] }
            }
          ],
          "type": "group",
          "name": "based-on-your-AddOn-Installed-SID"
        }
      ]
    }
  }
}
Log in to install VoiceBase Transcription w Custom Vocab
The Twilio advantage
  • Communicate reliably

    Experience a 99.95% uptime SLA made possible with automated failover and zero-maintenance windows.

  • Operate at scale

    Extend the same app you write once to new markets with configurable features for localization and compliance.

  • Many channels

    Use the same platform you know for voice, SMS, video, chat, two-factor authentication, and more.

  • No shenanigans

    Get to market faster with pay-as-you-go pricing, free support, and the freedom to scale up or down without contracts.