Classify viewer input

Classifies user input from the presentation viewer as either a system action (navigation, audio control, CTA trigger, continuation) or a question/statement. Uses GPT for classification with a regex-based fallback.

Authentication: No session required. Protected by IP-based rate limiting per presentation.

POST /api/actions

Classify viewer input

cURL

curl --request POST \
  --url https://app.demand-iq.com/api/actions \
  --header 'Content-Type: application/json' \
  --data '
{
  "text": "<string>",
  "presentationId": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
  "slideId": "<string>"
}
'

200

{
  "isAction": true,
  "targetSlide": "<string>"
}

Body

application/json

text · string · required

The user's input text

presentationId · string<uuid> · required

slideId · string

Current slide ID for context

Response

200

application/json

Classification result

isAction · boolean

Whether the input is a system action command

actionType · enum<string> | null

Available options: navigate_next, navigate_prev, navigate_slide, audio_pause, audio_resume, cta, continue, null

targetSlide · string | null

Target slide ID for navigation actions