Classify voice/text action

Classify user voice or text input as a presentation control command (navigation, audio, CTA) or route to Q&A. Uses AI classification with regex fallback.

POST /api/actions

Classify voice/text action

cURL

curl --request POST \
  --url https://app.demandiq.com/api/actions \
  --header 'Content-Type: application/json' \
  --data '
{
  "input": "Go to slide 3",
  "presentationId": "<string>",
  "context": {
    "currentSlide": 2,
    "totalSlides": 2,
    "isPlaying": true,
    "volume": 50,
    "awaitingResume": true,
    "didNavigateToNewSlide": true
  }
}
'

200

{
  "type": "<string>",
  "action": {
    "category": "navigation",
    "action": "<string>"
  },
  "confidence": 123
}

Body

application/json

input · string · required

Maximum string length: 500

Example:

"Go to slide 3"

presentationId · string

context · object

Show child attributes

Response

200 - application/json

Classification result

Option 1

Option 2

type · string

Allowed value: "action"

action · object

Show child attributes

confidence · number