YouTube template videos need distinct storylines: vary each brief

YouTube's channel monetization policies say channels cannot be mass-produced, generic or repetitive, and content produced using a template is at risk unless each video has a distinct storyline or focus. A bulk run that fills the same template with different product names does not meet that on its own. Give every item its own angle in the instruction, and check the batch for near-duplicates before you submit it.

What the policy page says

The monetization policies page, read on 2026-10-03, describes what YouTube treats as inauthentic content.

What a bulk run gives you and what it does not

The bulk runs docs describe a server-side queue of ordinary Format runs. The request takes a concurrency from 1 to 16 and an items array of 1 to 100 entries, each the same body as a single run. Each item must name at least one of instruction, input, previous_run_id or attachments.

Nothing in that contract makes two items different from each other. If forty items share one instruction apart from a product name, you get forty videos of one template. The variation has to be authored into the instructions.

Write the angle into each item

A distinct storyline means a different question, problem or story, not a different noun. Instead of repeating a review script for each product, assign each item a separate focus: a comparison against an older model, a use in a specific setting, a mistake people make, a repair, a size guide. The policy says storyline or focus, so each video should be describable in one sentence that no other video in the batch shares.

Build and screen the batch in Python

The script pairs each product with a different angle, rejects a batch where two instructions are too similar by word overlap, and submits the queue. The overlap check is a crude screen, not a policy test, and it only catches near-copies. The Format path is a placeholder to replace with yours, and the key must come from the environment.

import json, os, urllib.request, uuid
from itertools import combinations

def similar(a, b):
    wa, wb = set(a.lower().split()), set(b.lower().split())
    return len(wa & wb) / max(1, len(wa | wb))

briefs = [
    ("Trail shoe", "Compare how it grips wet rock against an older model"),
    ("Water bottle", "Show the one cleaning mistake that ruins the seal"),
    ("Day pack", "Pack it for a one-night trip and show what fits"),
]
items = [{"instruction": f"{p}: {a}"} for p, a in briefs]
for x, y in combinations(items, 2):
    if similar(x["instruction"], y["instruction"]) > 0.6:
        raise SystemExit("two items are too alike")

url = "https://api.sume.com/v1/formats/your-handle/your-format/bulk-runs"
req = urllib.request.Request(
    url, method="POST",
    data=json.dumps({"concurrency": 2, "items": items}).encode(),
    headers={
        "Authorization": "Bearer " + os.environ["SUME_API_KEY"],
        "Content-Type": "application/json",
        "Idempotency-Key": str(uuid.uuid4()),
    },
)
print(urllib.request.urlopen(req).read().decode())

After the queue is created

The docs say the queue has no webhook of its own and that completed means every item is terminal, not that all succeeded, so read counts.failed. Mint a fresh Idempotency-Key for each batch, because replaying a spent key returns the old queue. Poll status_url for progress.

Avoid the expert persona trap in the same pass: if a Format voices a presenter as a doctor, lawyer, adviser or political commentator, the policy line above applies to monetization. For the wider question of what the originality update changes, see the originality post.