Sentiment Analysis

sentiment boolean Default: false

Pre-recorded Streaming:Nova English (all available regions)

Deepgram’s Sentiment Analysis feature recognizes the sentiment of the entire transcript and detects shifts in sentiment throughout the transcript.

A sentiment of positive, negative, or neutral along with a sentiment_score is identified for each segment of the transcript, and an average of the entire transcript’s sentiment is also provided along with the average sentiment_score.

This sentiment feature produces sentiment analysis at every level of transcription requested, so you’ll get sentiment values for every word, sentence, utterance, and paragraph, in addition to the top-level sentiment for the entire transcription.

To enable Sentiment, use the following parameter in the query string when you call Deepgram’s /listen endpoint:

sentiment=true

To transcribe audio from a file on your computer, run the following curl command in a terminal or your favorite API client.

curl \
  --request POST \
  --header 'Authorization: Token YOUR_DEEPGRAM_API_KEY' \
  --header 'Content-Type: audio/wav' \
  --data-binary @youraudio.wav \
  --url 'https://api.deepgram.com/v1/listen?sentiment=true'

When the file is finished processing, you’ll receive a JSON response that has the following basic structure:

"metadata": {...},
  "results": {
    "channels": [
      {
        "alternatives": [...]
      }
    ],
    "sentiments": {
      "segments": [
        {
          "text": "Hi there. I love my current phone, but want a new one.",
          "start_word": 0,
          "end_word": 11,
          "sentiment": "positive",
          "sentiment_score": 0.4443286657333374
        },
        {
          "text": "Can I upgrade my phone?",
          "start_word": 12,
          "end_word": 16,
          "sentiment": "neutral",
          "sentiment_score": 0.1317894160747528
        }
      ],
      "average": {
        "sentiment": "positive",
        "sentiment_score": 0.35909067500721326
      }
    }
  }
}

The sentiment values added to objects are:

• sentiment: The sentiment of the associated text (‘positive’ | ‘negative’ | ‘neutral’).
• sentiment_score: A floating point value between -1 and 1 representing the sentiment of the associated span of text. -1 being the most negative sentiment, and 1 being the most positive sentiment.

If you request Sentiment Analysis with an unsupported language by specifying a language code such as sentiment=true&language=es or sentiment=true&detect_language=true where the detected language is unsupported, you will get the warning message below.

"warnings": [
  {
    "parameter": "sentiment",
    "type": "unsupported_language",
    "message": "Sentiment is only supported for English."
  }
]

Example Warning

Here is an example of the JSON structure of a request with warning object.

{
 "metadata": {
...
       },
     "warnings": [
       {
         "parameter": "sentiment",
         "type": "unsupported_language",
         "message": "Sentiment is only supported for English."
       }
     ]
   },
 "results": {
       "channels": [
            {
                "alternatives": [...]
            }
        ], 

    }
}