TikTok Shop Commerce Change Feed

Turn repeated TikTok Shop product checks into compact before-and-after evidence. Every submitted product produces a row, including a new baseline, an unchanged check, a detected change, a missing listing, or a source failure.

Accepted input

productUrls accepts TikTok Shop product page URLs and numeric product IDs. Use the same region and product list on scheduled runs so every observation is compared with the matching stored baseline.

{
  "productUrls": [
    "https://shop.tiktok.com/us/pdp/1730927783781307026",
    "1730949123456789012"
  ],
  "region": "US",
  "maxItems": 100
}

Response fields

Each row is one product check:

• product_id and product_input identify the checked listing.
• change_status is baseline_created, unchanged, changed, not_found, source_error, or invalid_product_reference.
• changed_count states how many tracked fields changed.
• changed_fields contains only the changes, each with field, before, and after values.
• observed_at is the current observation timestamp.
• analysis states whether a previous baseline was available.
• provenance records the product source and observation outcome.

{
  "record_type": "tiktok_shop_commerce_change",
  "product_id": "1730927783781307026",
  "product_input": "https://shop.tiktok.com/us/pdp/1730927783781307026",
  "change_status": "changed",
  "changed_count": 2,
  "changed_fields": [
    {
      "field": "sale_price",
      "before": "18.00",
      "after": "14.00"
    },
    {
      "field": "promotion_tags",
      "before": null,
      "after": ["flash sale"]
    }
  ],
  "observed_at": "2026-09-04T09:30:00.000Z",
  "analysis": {
    "status": "changed",
    "baseline_available": true
  }
}

Tracked evidence

The snapshot tracks title, seller ID, sale and original price, currency, stock status and quantity, promotion tags, rating, review count, and visible sold count when TikTok supplies them. Numeric formatting changes do not create a new observation.

changed_fields is empty for baseline_created and unchanged. A first run establishes the baseline and does not invent a change. Unavailable fields remain null rather than being inferred.

Scheduled runs

Use the same region and product list on scheduled runs. The first observation establishes a baseline; later observations compare each product with its previous successful baseline. Missing listings and collection failures leave that baseline unchanged.