Extend a Grok Imagine video on Sume: chain by last frame

xAI's guide lists video extension from the final frame. Sume's Grok row has no extend, so pull a late frame with video-frames and start the next clip from it.

5 min readSume
All posts

Sume cannot extend a Grok Imagine clip in one call, but you can chain clips yourself: extract a frame from just before the end of the first MP4, then submit it as the first_frame of the next Grok job. Each hop is a separate job and a separate charge, and 15 seconds is the most a single Grok job can run on Sume.

What xAI offers and what Sume exposes

xAI's video guide lists extension as a mode that continues a video from its final frame, alongside editing and reference-to-video. The Sume Grok row accepts one start image only, so extension is not offered. The chain below is a workaround, not the same feature, and the join is only as smooth as your prompt makes it.

Step 1: take a late frame

The video-frames API reads a clip stored on media.sume.com and returns durable stills at the times you list in at[]. Each time must satisfy 0 <= t < duration, so you cannot ask for the exact last instant. Ask for a point a fraction of a second before the end. If the clip is not already a media.sume.com artifact or asset of your workspace, import it first with POST /v1/media-imports, because the API refuses off-host URLs.

curl -X POST https://api.sume.com/v1/video-frames \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: chain-001-frame" \
  -d '{"video_url": "https://media.sume.com/artifacts/artf_demo/clip1.mp4", "at": [5.8], "format": "png"}'

Step 2: start clip two from it

When resource_status is ready, frames[0].url is a durable image. Send it as frame_images[0].image_url.url with frame_type: "first_frame" in the next POST /v1/videos for grok-imagine-video-1.5, and describe what happens next in the prompt.

Costs and caveats

Costs add up per hop at the Grok rate of 1.25 cents a second, plus the video-frames job, which Sume charges by its own compute time. Expect visible differences at the cut: Grok sees one image, not the motion before it, so keep the camera move and lighting words the same across prompts and join the clips on a cut or a short transition in the timeline.

Chaining three 10-second Grok clips on Sume (read 2026-10-13)
HopSecondsUnrounded Grok cost
Clip 110$0.125
Clip 2 from a frame of clip 110$0.125
Clip 3 from a frame of clip 210$0.125
Total video30$0.375 unrounded; $0.39 billed (3 x $0.13)

Sources

Related posts

More in Developers

All Developers posts

Written by Sume