Wan 3.0 result URLs expire in 24 hours at Alibaba; fetch from Sume

Alibaba Cloud's Wan3.0 guide says the result URL is valid for 24 hours, so you have to download the file in that window (read 2026-10-04). If your pipeline calls Wan 3.0 directly, a late download means a re-run and a second charge.

Through Sume you do not hold the vendor URL. A finished job exposes its video at GET /v1/videos/{jobId}/content?index=0, and you send your Bearer key to read it, per the video generation docs. This post does not claim how long Sume keeps the file, because the docs do not put a number on it here. Download what you need as soon as the job is done and store it yourself.

Download on completion

import os
import requests

job_id = os.environ["JOB_ID"]
url = f"https://api.sume.com/v1/videos/{job_id}/content?index=0"
r = requests.get(
    url,
    headers={"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"},
    stream=True,
)
r.raise_for_status()
with open(f"{job_id}.mp4", "wb") as f:
    for chunk in r.iter_content(1 << 20):
        f.write(chunk)
print("saved", job_id)

A safer pipeline

Cost is unaffected by the download, so retrying a fetch is free. A failed job is the case to watch instead; check the job status before you assume a clip exists.

• Pass a callback_url on submit if you want a push when the job finishes, and download in that handler.
• Save the MP4 to your own storage on first read.
• To edit it further, import it to media.sume.com and use Timeline or Video trim, which only accept Sume-hosted URLs.