Assemble a three-clip montage with fades through the Sume API

Post one audio spine and an ordered video[] list to /v1/timeline-1.0/render: three clips with fade and dissolve transitions over a 30-second spine render as one MP4 for $0.10 (Timeline 1.0). The first slot must start at 0, later starts must increase, and every URL must already be on media.sume.com.

Python request

Python has no top-level await, so this is plain synchronous code. The job is async by default; poll the job envelope.

import os, time, requests

API = "https://api.sume.com"
h = {"Authorization": "Bearer " + os.environ["SUME_API_KEY"]}
m = "https://media.sume.com/artifacts/artf_demo/"
body = {
    "audio": {"url": m + "music.wav", "duration_seconds": 30},
    "video": [
        {"source_url": m + "a.mp4", "start": 0, "duration": 10},
        {"source_url": m + "b.mp4", "start": 10, "duration": 10,
         "transition": {"type": "fade", "duration": 0.5}},
        {"source_url": m + "c.mp4", "start": 20, "duration": 10,
         "transition": {"type": "dissolve", "duration": 0.5}},
    ],
}
r = requests.post(API + "/v1/timeline-1.0/render",
                  headers={**h, "Idempotency-Key": "montage-001"}, json=body)
r.raise_for_status()
j = r.json()
jid = j.get("data", j)["request_id"]
while True:
    s = requests.get(API + "/v1/jobs/" + jid + "/status", headers=h).json()
    s = s.get("data", s)
    if s["terminal"]:
        break
    time.sleep(s.get("next_poll_after_seconds") or 3)
res = requests.get(API + "/v1/jobs/" + jid + "/result", headers=h).json()
print(res.get("data", res))

Rules the compiler enforces

Frame rate

Leave output.fps out unless you need it. If an explicit rate differs from your sources the job resamples and reports output_fps_resamples_sources.

Check the plan first

POST /v1/timeline-1.0/plan takes the same body and returns duration_seconds, segment_count, billable_minutes and estimated_cost_usd_micros without creating a job or reserving credits, and it does not need an Idempotency-Key. For this request the plan should show 30 seconds, 3 segments and 1 billable minute. A plan cannot predict short-source pad or loop warnings, so read warnings[] on the finished result as well.