Text Intent Recognition

The Intent Analysis model will help you capture semantics behind users' messages and assign it to the right label. It is designed to understand human conversation in the form or free text or spoken text.

API Request

Returns the predicted intent for a given phrases / sentences

POST https://api.marsview.ai/text/intent/get_intent

Query Parameters

Name	Type	Description
jwtToken*	String	Marsview api access token. You can generate it here: https://docs.marsview.ai/speech-analytics-api/authentication
sentences*	List	The sentences you want to match intents for
userId*	String	Your Email or User ID
threshold	Float	The confidence threshold for intent matching. If not specified defaults to 0.5
intentIds*	List	List of intent ids to match the sentences against

200: OK Response with recognized intents

{
  "status": true,
  "error": {},
  "data": {
    {
    "similarityMatch": [
        {
          "matchedTextList": [
            {
              "score": 0.6180883955955505,
              "text": "Marsview is a great place to work",
              "value": "Marsview"
            }
          ],
          "intentId": "intent-bxllq2f7hpkrvtyzi3-1627981197627",
          "intentName": "Marsview"
        },
      ],
    }
  },
  "userId": "[e-mail]"
}

400: Bad Request When a bad request is sent with missing key

{
    "status": false,
    "error": {
        "message": "Bad Request Missing key",
        "code": "RTSINT002"
    },
    "data": {},
    "userId": "[e-mail]"
}

401: Unauthorized When the user requested with an invalid jwt token

{
    "status": false,
    "message": "Something went wrong while validating your JWT token."
}

Example Curl Request

Shown below is a code snippet to send a sample request for intent recognition

curl --location --request POST 'https://api.marsview.ai/text/intent/get_intent' \
--header 'Content-Type: application/json' \
--data-raw '{
    "jwtToken" : "Insert JWT token Key",
    "threshold": "Insert a desired confidence threshold",
    "intentIds": "Insert a list of intent IDs", 
    "sentences": "Insert your sentence for intent recognition",
    "userId"   : "[e-mail]"
}'

Example Response

Shown below is a sample JSON response from the above request.

{
  "status": true,
  "error": {},
  "data": {
    {
    "similarityMatch": [
        {
          "matchedTextList": [
            {
              "score": 0.6180883955955505,
              "text": "Marsview is a great place to work",
              "value": "Marsview"
            }
          ],
          "intentId": "intent-bxllq2f7hpkrvtyzi3-1627981197627",
          "intentName": "Marsview"
        },
      ],
    }
  },
  "userId": "[e-mail]"
}

Response Object

Field	Description
similarityMatch	The intent similarity output from the models
intentId	The intent Id against which the sentences are matched
intentName	The name of the matching intent
score	The similarity matching score ranges from 0 to 1
text	The sentence which is used to predict intent for
value	The predicted intent