Grok Imagine Video 1.5 API in Python: animate a still on Sume

To animate a still with Grok Imagine Video 1.5 on Sume, POST model: "grok-imagine-video-1.5" to /v1/videos with the image as a first_frame in frame_images. A 12-second 720p clip costs $0.15.

This row is image-to-video only: a text-only request is rejected, and so is any input_references, end frame, aspect_ratio or generate_audio.

The script

Needs pip install requests and SUME_API_KEY.

import os, time, requests

H = {"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"}
BASE = "https://api.sume.com/v1/videos"
body = {
    "model": "grok-imagine-video-1.5",
    "prompt": "Soft wind moves the curtains, the cat blinks slowly",
    "duration": 12,
    "resolution": "720p",
    "frame_images": [{
        "type": "image_url",
        "image_url": {"url": "https://example.com/cat-window.jpg"},
        "frame_type": "first_frame",
    }],
}
r = requests.post(BASE, headers=H, json=body, timeout=60)
print(r.status_code, r.text)
job = r.json()
while r.status_code == 202:
    s = requests.get(job["polling_url"], headers=H, timeout=60).json()
    if s["status"] in ("completed", "failed", "cancelled"):
        print(s["status"], s.get("unsigned_urls"), s.get("usage"))
        break
    time.sleep(10)

The limits

The numbers below come from the Grok row in the Video Router catalog and the provider pricing table.

When it fits

Grok is the lowest-priced row in the catalog, so use it for loops, product tiles and previews where silence is fine. Twelve seconds costs $0.15, and the 15-second maximum costs $0.19.

If you need sound, a text-only prompt or a last frame, pick another row: Gemini Omni Flash 1.1 for sound at 3 to 10 seconds, Wan 3.0 for a 2 to 30 second range, Kling 3 for a first and last frame. The GET /v1/videos/models endpoint lists which each row supports.

• Poll polling_url; download from unsigned_urls[0].
• Send an Idempotency-Key on retried submits.
• Source images must be publicly reachable HTTPS URLs.