Video

4 min

Nima allows you to automatically detect harmful videos using the AI detection providers you have configured.

Example payload

{
  "priority": "severe",
  "channel": "automated detection",
  "reason_for_request": "Lorem Abuse reason_for_request",
  "customerSpecific": {
  },
  "content": {
    "content_id": "098765434567",
    "title": "Sniper video",
    "body": "Forest walk holding weapon",
    "url": "https://videos.pexels.com/video-files/6201055/6201055-uhd_3840_2160_25fps.mp4",
    "token": "",
    "type": {
      "id": "VIDEO_STATIC"
    },
    "status": "published",
    "customerSpecific": {}
  },
  "reportee": {
    "id": "123",
    "name": "Reportee John",
    "lang": "en",
    "phoneNumber": "070-564123",
    "emailAddress": "[email protected]",
    "customerSpecific": {}
  },
  "label": [
    "WEAPON"
  ]
}

Request payload requirements

For video moderation, ensure your request payload includes the following content-type specific requirements:

Attribute	Required	Description
content.url	Yes	For video-based moderation, the URL should point to the video. If the URL provided is intended to display an entire page, please see "Detecting Harmful Pages"
content.type.id	Yes	Must be set to "VIDEO_STATIC"

Configuring detection rules

To create a rule for automated video detection, navigate to Configure > Trust & Safety > Rules and define the following conditions:

Required Conditions:

Channel: Must be set to "Proactive Detection"

Content Type: Must be set to "Video Static"

Once these conditions are set, the list of available AI detection providers will be accessible via the Choose Provider button.

Audio profanity detection

Sightengine supports audio profanity detection for video content (VIDEO_STATIC type). This allows you to moderate both the visual and audio components of videos in a single request.

When configuring a rule for video content with audio profanity detection:

Set the Content Type to VIDEO_STATIC

Select Sightengine as the provider

Choose the audio-profanity class

The detection processes synchronously and returns profanity matches found in the video's soundtrack along with a transcription. Videos without sound are handled gracefully—the system processes the content normally without the audio profanity component.