Four Twitch clips into one 60-second recap: trim, caption, join

Trim each clip, caption it, detach its audio, and join the four on Timeline 1.0 with fit: "blur" so landscape footage fills a vertical frame. Four 15-second clips become a 60-second recap for $1.02 on Sume, with no generation involved.

The chain

Start from clip files you own the rights to and import them into the workspace (POST /v1/media-imports). Trim accepts a start and a duration of 0.2 to 900 seconds. Captions need audible speech or they fail with caption_no_speech, so send cues for a silent moment.

curl -X POST https://api.sume.com/v1/video-captions \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: recap-cap-1" \
  -d '{
    "video_url": "https://media.sume.com/artifacts/artf_demo/clip1-trim.mp4",
    "style": "punch",
    "language": "en"
  }'

The join

Use the detached audio of each captioned clip as audio.parts[] slices, in order, so speech is gapless across cuts. The four video slots use blur fit with a short fade between them.

{
  "audio": { "parts": [
    { "url": "https://media.sume.com/artifacts/artf_demo/a1.wav", "duration": 15 },
    { "url": "https://media.sume.com/artifacts/artf_demo/a2.wav", "duration": 15 },
    { "url": "https://media.sume.com/artifacts/artf_demo/a3.wav", "duration": 15 },
    { "url": "https://media.sume.com/artifacts/artf_demo/a4.wav", "duration": 15 } ],
    "duration_seconds": 60 },
  "video": [
    { "source_url": "https://media.sume.com/artifacts/artf_demo/v1.mp4", "start": 0, "duration": 15, "fit": "blur" },
    { "source_url": "https://media.sume.com/artifacts/artf_demo/v2.mp4", "start": 15, "duration": 15, "fit": "blur",
      "transition": { "type": "fade", "duration": 0.25 } }
  ]
}

Cost per recap

The caption price is per job for videos up to 60 seconds, so each 15-second clip is one job.

Weekly math

One recap a week for a year is $53.04. Extra clips add $0.23 each.

Order and pacing

Put the strongest moment first and the funniest last, and keep the fade at a quarter of a second so the join does not eat the first word of the next clip. Captions are burned into each clip before you join them, so a caption style choice applies per clip; pick one style for the whole recap so the text does not change look between cuts. Always use footage you have the right to reuse, and credit the people in it.