TikTok Shop Good-tier listings: five images over 600x600, batched

TikTok Shop's seller guidance puts a listing in the Good quality tier when it has five or more images, each above 600x600 pixels, with a clean main image. If you generate the set with an image API, plan five distinct shots per product, not five crops of one picture, and check the pixel size of every file you download.

What the seller page lists

The TikTok Shop seller page on listing quality describes tiers for listing quality. These are the items that matter for an image batch, as read on 2026-10-03.

Plan the set as a shot list

Five images only count if they show different things. A shot list written before generation keeps a batch from producing five near-identical frames. One workable split for a single product:

• Image 1: the clean main image on a white background, nothing else in frame.
• Image 2: the product from a second angle.
• Image 3: a detail or material close-up.
• Image 4: the product with something that shows its size.
• Image 5: the product in use, in a lifestyle setting.

Batch it with the Sume Image API

The Image API takes a model id, a prompt, and optional fields such as aspect_ratio and resolution, which each model lists in supported_parameters. A request that exceeds the 30-second budget returns 202 with a job envelope instead of the image, so the script below handles both. Read each model's supported_parameters from GET /v1/images/models before pinning a ratio or tier.

For a catalog with many products, a Format run queue is the other route: bulk runs accept 1 to 100 items with a concurrency window of 1 to 16.

import os
import requests

URL = "https://api.sume.com/v1/images"
HEADERS = {
    "Authorization": f"Bearer {os.environ['SUME_API_KEY']}",
    "Content-Type": "application/json",
}
SHOTS = [
    "Product on a plain white background, studio light, nothing else in frame",
    "Same product, three-quarter angle, plain white background",
    "Close-up of the stitching and material, soft light",
    "Product beside a hand for scale, neutral surface",
    "Product in use on a kitchen counter, daylight",
]

for i, shot in enumerate(SHOTS, start=1):
    body = {
        "model": "bytedance-seed/seedream-4.5",
        "prompt": shot,
        "aspect_ratio": "1:1",
    }
    r = requests.post(URL, headers=HEADERS, json=body, timeout=60)
    if r.status_code == 200:
        print(i, r.json()["data"][0]["url"])
    elif r.status_code == 202:
        print(i, "queued:", r.json()["data"]["status_url"])
    else:
        print(i, "error", r.status_code, r.text[:200])

Check pixel size on the files you keep

When a model accepts it, Sume's resolution field is a normalized tier (512, 1K, 2K or 4K), not a promise of an exact pixel count. Open each downloaded file and read its width and height before upload. Anything at or below 600 on a side does not meet the page's resolution line, and a gallery with four qualifying images is still short of the five the tier asks for.

Also keep the main image honest. The page excludes watermarks, borders, promo text and logos from it, so leave those out of the main-image prompt and add any branding to later images only if you want it there.