Shopify Review Rating & Count Delta Monitor

Product-bound public Judge.me widget rating/count snapshots with optional caller-supplied baseline comparison. No reviewer names, text, photos or contact details are exported.

Quick start

Supply 1-10 canonical public Shopify product URLs. The prefilled example processes one product. No Shopify login, API key, browser, proxy or AI service is needed.

Example input:

{"productUrls":["https://norlanglass.com/products/norlan-whisky-glass"]}

Optional baselines: a JSON object keyed by the exact product URL. Each value contains productId (numeric string), averageRating (number 0-5), reviewCount (nonnegative integer), observedAt (past ISO timestamp). The same URL and product ID must match. Each run is stateless; copy prior report metrics into your next input using your own scheduler. Unsupported products are listed in OUTPUT without charge.

Results

One dataset row per successfully processed product; nested arrays stay structured in JSON. Use Apify dataset exports for JSON/CSV/Excel. Common fields: stableKey, productId as string, productUrl, sourceUrl, productTitle, checkedAt. Specific fields: averageRating, reviewCount, baselineAverage, baselineCount, ratingDelta, countDelta, widgetUpdatedAt, provider, association. OUTPUT reports per-product source errors, stored report count, charged count and budget truncation.

Pricing

Pay per event USD0.004 per saved product report (USD4 per 1,000 reports). This is per product report, not per asset, finding, review or variant. Failed/unsupported sources and diagnostics are free. A valid report with zero findings, zero reviews or zero media can be chargeable because the requested check succeeded. Platform usage charges may also apply. Every report is stored before its event is charged. Storage and billing are separate services; interruption between them is not an atomic transaction. A billing/storage failure stops processing rather than silently continuing.

Scope and limitations

Only server-rendered Judge.me widgets whose data-id exactly matches the public Shopify product ID. No multi-provider or JavaScript rendering claim. Missing widgets are UNSUPPORTED and free. A widget may aggregate grouped products; metrics describe the widget, not verified SKU-specific reviews or purchases. Counts may fall after moderation. A rating delta is not sentiment analysis or evidence of fraud. Baseline is caller-supplied, not authenticated historical truth; no persistent named state or automatic baseline updates.

Each run reads at most 10 selected products, 20 primary HTTP requests (one product JSON and one HTML per product), at most 3 same-origin redirects per request,4MiB per response and 120 seconds of source work. No unbounded pagination or retries. Headless/password-protected stores may be unsupported; public source formats may change. All unsupported/failing inputs fail the run when no supported report can be produced; partial success is explicit in OUTPUT. Default QA input is small and intended to finish under5 minutes. Product IDs and handle/source URL are verified; unsafe numeric IDs are rejected rather than rounded.

Security and privacy

Read-only public HTTPS with DNS/IP validation and pinning, TLS hostname verification, same-origin redirects, hard caps and no arbitrary scripts, cookies, login or private network access. Product metadata may contain merchant-provided content; no customer/reviewer personal information is intentionally exported. No Admin API or cart writes. Not affiliated with Shopify or Judge.me. Respect content rights and target terms.

Development

Run npm ci, npm test and apify validate-schema. See TEST_PLAN.md and BUILD_PUBLISH_CHECKLIST.md for acceptance. Pricing is configured only for this new Actor; no changes to existing public Actor prices.