Search Videos | Coactive AI

curl -X POST https://api.coactive.ai/api/v1/search/video \
     -H "Authorization: Bearer <token>" \
     -H "Content-Type: application/json" \
     -d '{
  "dataset_id": "12345678-abcd-90ef-1234-56789abcdef0",
  "text_query": "President speaking at a podium",
  "offset": 0,
  "limit": 60
}'

{
  "results": [
    {
      "dataset_id": "12345678-abcd-90ef-1234-56789abcdef0",
      "text_query": "President speaking at a podium",
      "video_id": "b8b2b7a6-7d2a-4a0c-8c7a-93b0b2a3e111",
      "source_path": "s3://demo-bucket/videos/2024/clip_001.mp4",
      "video_score": 0.2,
      "metadata": {
        "data": {
          "headline": "President addresses reporters at White House",
          "caption": "The President delivers remarks from a lectern in the East Room.",
          "location": "East Room"
        }
      },
      "top_composite_slice": {
        "id": "5f7f9c4a-2a4f-4f4e-bf3b-5b6e8f6de001",
        "index": 3,
        "start_frame_num": 1530,
        "end_frame_num": 2100,
        "start_time_ms": 51000,
        "end_time_ms": 70000,
        "composite_id": "4b0a3c9f-5f2d-4b1e-8b0a-0a9f4b1e2c33",
        "composite_type": "shot",
        "composite_slice_score": null
      },
      "top_keyframe": {
        "id": "c2b0e1f2-8899-4f1d-9a66-6df1bb9b77aa",
        "status": null,
        "keyframe_score": null,
        "start_time": null,
        "end_time": null
      }
    }
  ]
}

Search your dataset with natural language and return one result per video, surfacing the most relevant composite slice (top_composite_slice) and a preview keyframe.

A composite is how Coactive segments a video into meaningful parts. Entries in responses are composite slices with start_time_ms and end_time_ms. Unlike /api/v1/search/text-to-image, which returns keyframes, this returns videos and their best-matching moments.

Note: This works only for videos ingested with Coactive’s new video data model. If you’re unsure or want to try it, contact Coactive Support.

Search your dataset with natural language and return one result per video, surfacing the most relevant composite slice (top_composite_slice) and a preview keyframe. A composite is how Coactive segments a video into meaningful parts. Entries in responses are composite slices with start_time_ms and end_time_ms. Unlike /api/v1/search/text-to-image, which returns keyframes, this returns videos and their best-matching moments. Note: This works only for videos ingested with Coactive’s new video data model. If you’re unsure or want to try it, contact Coactive Support.

Authentication

AuthorizationBearer

Bearer authentication of the form Bearer <token>, where token is your auth token.

Request

This endpoint expects an object.

dataset_idstringRequiredformat: "uuid"

text_querystringRequired>=2 characters

Text query must be at least 2 characters long

metadata_filtersobject or nullOptional

offsetintegerOptionalDefaults to 0

limitintegerOptionalDefaults to 60

Response

Successful Response

resultslist of objects

Authentication

Request

Response

Errors