Four stills to one AI video: three H3 transitions, $1.225 total

The short answer

To turn four stills into one video, request three clips where each uses one still as first_frame and the next as last_frame: A to B, B to C, C to D. Three 5-second MiniMax H3 clips at 768p cost $1.125 (15 seconds at $0.075). Joining them in a Timeline render adds $0.10, for $1.225 in all.

Why three clips

Four photos give three gaps, so you need three clips. Each clip is its own /v1/videos job. H3 accepts 5 to 15 seconds and supports an end frame, so each job takes two images and invents the move between them.

Clip A to B

The first job below goes from still A to still B. The second uses B and C, and the third C and D.

curl -X POST https://api.sume.com/v1/videos \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: chain-ab" \
  -d '{
    "model": "minimax-h3",
    "prompt": "Smooth cinematic move from the first scene to the second",
    "duration": 5,
    "resolution": "768p",
    "frame_images": [
      {"type": "image_url", "image_url": {"url": "https://example.com/a.png"}, "frame_type": "first_frame"},
      {"type": "image_url", "image_url": {"url": "https://example.com/b.png"}, "frame_type": "last_frame"}
    ]
  }'

Join the clips

Because the last frame of each clip is the first frame of the next, the joins land on the same picture. Keep the cuts hard or use a short fade. A fade of 0.25 seconds is allowed on every slot after the first, and it must be no more than 50% of the shorter neighbor.

Timeline render

Import the three finished clips to the media host, then render. The audio spine can be silence at the declared length of 15 seconds, and if you want music you can add a soundtrack bed.

curl -X POST https://api.sume.com/v1/timeline-1.0/render \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: chain-join" \
  -d '{
    "audio": {"mode": "silence", "duration_seconds": 15},
    "video": [
      {"source_url": "https://media.sume.com/artifacts/artf_demo/ab.mp4", "start": 0, "duration": 5},
      {"source_url": "https://media.sume.com/artifacts/artf_demo/bc.mp4", "start": 5, "duration": 5},
      {"source_url": "https://media.sume.com/artifacts/artf_demo/cd.mp4", "start": 10, "duration": 5}]
  }'

What happens to the clip audio

H3 audio is always on, and a silence spine replaces it. To keep the clip sounds, use the clips' detached audio as the spine instead (see Audio detach).

Total

The prices are billed at list x 1.25, read from the Sume repo on 2026-10-05.