A video in. Its breakdown out.

The API turns a video into data a model can work from: every word with its time, every cut and what kind it is, every caption and graphic with its box, and who is in it. Keys, jobs, credit and receipts live here.

{
  "format": "video-analysis", "version": "1",
  "video": { "name": "clip.mp4", "duration": 13.5,
             "width": 720, "height": 1280 },
  "transcript": { "language": "en", "words": [
    { "w": "This", "s": 3.52, "e": 3.71 } ] },
  "cuts": [ { "t": 3.5, "kind": "cut" } ],
  "shots": [ { "n": 2, "s": 3.5, "e": 6.73,
               "scene": 2, "kind": "cut" } ],
  "elements": [ { "kind": "caption",
    "text": "This effortless new way",
    "s": 3.5, "e": 4.75,
    "box": { "x": 0.14, "y": 0.62,
             "w": 0.72, "h": 0.05 } } ],
  "characters": { "cast": [ { "label": "A",
    "name": "Golden retriever", "role": "pet" } ] }
}