Product JSON-LD Normalizer — Offers, Prices & Variants

This Apify Actor fetches one public static HTML page and stores exactly one deterministic page result containing normalized Schema.org Product, ProductGroup, Offer, and AggregateOffer JSON-LD.

Input

{
  "url": "https://example.com/product/sku-1",
  "timeoutSeconds": 15,
  "maxResponseBytes": 5242880,
  "maxRedirects": 3
}

Only url is required. Unknown fields and non-public destinations are rejected. Every redirect is resolved and SSRF-checked again. The fetcher pins the connection to the checked IP and bounds both compressed and expanded bodies.

Output

A successful fetch pushes exactly one item to the default dataset, including pages with no Product or malformed JSON-LD. Missing values remain null or empty arrays. Prices are parsed without a binary float and emitted as non-exponent decimal strings. Every normalized node and finding includes a deterministic sourcePath.

Finding codes cover malformed JSON-LD, missing products and identifiers, absent or invalid offers, price/currency/availability problems, invalid aggregate ranges, ambiguous scalar fields, and unresolved references.

Supported scope

• One public HTTP(S) URL and static HTML only
• application/ld+json top-level objects, arrays, @graph, and embedded hasVariant
• Product and ProductGroup nodes with embedded Offer or AggregateOffer
• @type as a string or array

The Actor does not render JavaScript, browse or crawl, use proxies, parse Microdata/RDFa, resolve external @id values, inherit group data into variants, validate live stock, or store raw HTML/JSON-LD. It needs no Secret, API key, paid API, browser, or external service.

Local verification

From this directory with the repository virtual environment available:

..\..\.venv\Scripts\python.exe -m unittest discover -s tests -v
..\..\.venv\Scripts\python.exe -m ruff check .

For a local Actor run, put input in storage/key_value_stores/default/INPUT.json, expose only a local fixture through the test resolver or use an explicitly approved public page, and run apify run. Local tests use fixture HTTP only and cost ¥0.

Release boundary

Human Owner approved Product #2 Public release, Monetization, and $0.002 per successful apify-default-dataset-item on 2026-08-15. Platform activation remains pending candidate build and production verification; this Release Candidate source artifact alone is not a declaration that those settings are active. Paid cloud runs and external spending remain separate approval-gated actions.