Product Data for AI Shopping Agents

Extracts e-commerce product data and normalizes it into one consistent, null-safe schema — built for feeding AI shopping agents, RAG pipelines, and price/catalog pipelines a data source they can actually trust.

The problem this solves

Every e-commerce site structures its product pages differently. An AI agent (or a script) trying to read raw HTML across multiple retailers gets inconsistent results — different field names, missing data, broken parsing on every redesign. This Actor reads structured data instead of guessing from layout: schema.org/JSON-LD first (what most SEO-conscious stores already embed for Google), falling back to Open Graph tags when JSON-LD isn't present.

Input

• Product page URLs — a list of URLs to process
• Max pages to process — safety limit per run (default 100)

Output

One record per product, always the same shape:

{
  "id": "SKU123",
  "name": "Product Name",
  "brand": "Brand",
  "price": 29.99,
  "currency": "USD",
  "availability": "InStock",
  "rating": 4.6,
  "reviewCount": 128,
  "url": "https://example.com/products/...",
  "imageUrl": "https://example.com/img.jpg",
  "description": "...",
  "embeddingText": "Name — Brand — Description",
  "source": "json-ld",
  "scrapedAt": "2026-09-11T04:01:52.791Z"
}

source tells you which extraction path fired — "json-ld" (richer: includes brand, rating, reviews) or "og-fallback" (thinner, but still usable — every field is present, explicitly null when unavailable, so you never have to defensively check for missing keys).

Why JSON-LD first

Reading structured markup the site already maintains for Google's own crawler means far less breakage than scraping raw DOM — no per-site CSS selectors to fix every time a theme changes.

Using it with an AI agent

Point an agent's tool call at this Actor with a list of product URLs and get back clean, structured data ready to embed, compare, or reason over — no HTML parsing on the agent's side.

Pricing

Pay-per-event: charged per product record returned, not per page crawled or per run. No charge for pages that don't contain product data.

Known limitation

Works on pages that carry schema.org/JSON-LD or Open Graph product markup. Sites that render everything client-side in JavaScript without either will return no data — there's no headless browser in this version.