Merchant Feed QA — Price, Stock & Product Data Audit

Check your product feed before the next campaign or promotion. This Actor validates Google-style feed fields and compares valid rows with the price, currency and availability in the product landing page's server-rendered JSON-LD.

Built for shopping-ad specialists, ecommerce operators and agencies. Output is a prioritized issue list with evidence and a suggested fix for each finding.

Start here

For a fictional demonstration with no web traffic or Actor event fees:

{"demoMode": true}

For a real check, replace these example URLs with your own:

{
  "products": [{
    "id": "bag-blue",
    "sku": "BAG-BLUE",
    "title": "Everyday Backpack",
    "link": "https://your-store.example/products/backpack",
    "image_link": "https://your-store.example/images/backpack.jpg",
    "price": "79.00 USD",
    "availability": "in_stock"
  }],
  "maxItems": 100
}

Alternatively supply feedUrl for a public CSV, TSV, JSON, RSS or Atom feed. Do not supply both products and feedUrl. No Merchant Center account connection is required. The .example URLs above are placeholders, not live targets.

What you receive

• Feed checks: missing essential fields, malformed URLs, invalid price/currency, invalid sale intervals, unsupported availability and invalid GTIN checksum.
• Landing-page checks: price, currency and stock differences; confirmed 404/410.
• Change comparison: new and resolved issues against an earlier snapshot.
• Dataset rows with issues, data, changes, status and billable.
• SUMMARY, SNAPSHOT, REPORT and a formula-safe EXPORT CSV in run storage.

Feed price format is "19.99 USD". Optional sale_price follows the same format. sale_price_effective_date is a timezone-aware ISO interval such as "2026-11-20T00:00:00Z/2026-11-30T00:00:00Z". Without an interval, a valid sale price is considered active. A sku selects a particular Product/Offer when present.

Pricing

Proposed launch price: $5 per 1,000 audited products ($0.005 each), with normal Actor runtime costs included. No startup fee or monthly rental. Billing event: product-audited. Price must be activated in Apify Publication before sale.

A completed audit that discovers invalid feed fields is billable even when no page fetch is necessary. Confirmed missing landing pages are also useful, billable findings. Network failures, blocks and ambiguous extraction are unbillable. Duplicate feed IDs are skipped with a summary notice. One product is billed once per run, regardless of its number of findings. Unchanged audits are still checks. Set the Apify maximum charge per run to control spending. A 100-product audit costs at most $0.50 in Actor event fees.

Monitoring

Paste the previous SNAPSHOT object into previousSnapshot, or supply an existing stateStoreId you own together with a unique monitorId. The latter requires explicit storage permissions on Apify. Schedule non-overlapping runs for the same monitor. Store separate customers' baselines in separate stores. Unverified rows retain their previous observations. The most recent 5,000 observations are retained.

The first observation establishes the baseline. This Actor returns changes; it does not send messages. Use your own Apify/n8n/Make integration to route findings.

Scope and limits

This is a focused diagnostic, not a complete Google policy validator, a feed editor or a guarantee of approval. It compares JSON-LD, not the rendered checkout, tax, shipping, discounts requiring login, geographic personalization or all visible prices. JavaScript-only data, AggregateOffer ranges and ambiguous multiple offers are not guessed. Missing availability is a warning, not a match.

Rows with invalid feed fields receive a validation-only audit; fix them and rerun to inspect their landing pages. GTIN checks validate the checksum, not ownership. Image URL syntax is checked; image files are not downloaded.

Maximum 1,000 unique items/run; 2 MB per HTTP response after decompression; 10-minute crawl ceiling; bounded redirects and retries. Only public HTTP(S) pages and ports 80/443 are supported. robots.txt is respected. No proxies, browser rendering, credentials or paid third-party APIs are required. Access failures are reported. Use only feeds and pages you have permission to analyze.

Apify account plans, optional external integrations, and dataset access after the run may incur separate platform charges. The demo has no Actor event fees; platform execution can still consume usage.