Cost of one AI video job: read GET /v1/usage with job_id

To get the cost of one AI video job on Sume, call GET /v1/usage?job_id=<id> and quote summary.debited_usd. That field is what the wallet deducted for the job. Do not add up ledger rows yourself: a refunded row keeps its hold amount in billable_amount_usd_micros, so a hand-made sum overstates spend.

Video prices change with each launch week, so a per-job lookup is more reliable than a spreadsheet of assumed rates. The behavior below is from Sume's usage docs, read 2026-10-02.

What does the summary contain?

The response adds a summary folded over every ledger row the job caused. The limit parameter only caps the rows listed, not the summary.

How do the statuses show up?

Rows move through reserved, captured and refunded. A video job reserves an estimate on submit, captures billable usage after success, and releases the reservation after a failure or cancellation before capture. Sume bills provider list times 1.25 on video models, and usage.cost in the poll response is the Sume billable amount.

So read final first. While it is false, a hold is still open and the debit may change.

import os
import requests

r = requests.get(
    "https://api.sume.com/v1/usage",
    params={"job_id": "JOB_ID_HERE", "limit": 50},
    headers={"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"},
    timeout=30,
)
r.raise_for_status()
s = r.json()["summary"]
print("debited:", s["debited_usd"], "final:", s["final"])

When should I use this?

Use it after a test clip to record the real cost next to the prompt, and after a failed job to confirm the hold was released. Replace JOB_ID_HERE with the id from the submit response. For a batch, loop over your saved job ids and total only the debited_usd values, once each job reports final.