Luma modify video: ray-flash-2 15 seconds, ray-2 10, 100 MB

Luma's Modify Video endpoint limits the source clip by model: ray-2 takes up to 10 seconds and ray-flash-2 up to 15 seconds, and both cap the file at 100 MB. On Sume, video editing is the video_url mode of gemini-omni-flash-1.1, where a prompt describes the edit and you never pick an endpoint.

Luma's contract

Luma's page requires four fields: prompt, media.url, mode and model. An optional first_frame.url is preferred as a guide for the desired first frame. The two models are ray-2 and ray-flash-2.

Sume's edit mode

In edit mode the Video Router docs say the prompt describes the edit, resolution is optional with a default of 720p, and there is no aspect_ratio or duration. video_url is the edit source, not a reference, so it cannot be combined with image_url, end_image_url or reference_*_urls. Native audio is always on for this model.

curl -X POST https://api.sume.com/v1/video-router/generate \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: video-router-edit-001" \
  -d '{
    "model": "gemini-omni-flash-1.1",
    "prompt": "Replace the bottle with an apple. Keep everything else the same.",
    "video_url": "https://example.com/clip.mp4",
    "resolution": "720p",
    "mode": "async"
  }'

Check your own clips

Because the Sume pages read here do not state a source-length or file-size limit for edit mode, test with a real clip and read the error before you build a pipeline that relies on a length. Trim long sources first, and keep your own source files under Luma's 100 MB if you plan to use both.

Where to confirm limits

The model catalog is the place to confirm what a model accepts. Read capabilities from GET /v1/video-router/models rather than assuming one envelope.