JSON-LD Product Price & Stock Monitor

Check public product pages for price and stock changes by reading the schema.org Product JSON-LD already on the page.

This is not a marketplace searcher and not an Amazon/Walmart scraper. You supply the product URLs. The Actor reads public structured data only. It does not log in, does not bypass bot checks, and does not open hidden inventory APIs.

Who pays for this

• Brands watching their own and competitor public product pages
• Agencies running weekly price/stock checks on a known URL list
• Agents that need a small, typed offer record (price, currency, availability, changeType)

Repeat usage is the point: schedule the same URL list. The first successful check stores a snapshot. Later runs emit price_up, price_down, back_in_stock, out_of_stock, or unchanged.

Input

Minimum: one or more public product-page URLs.

{
  "startUrls": [
    { "url": "https://www.glossier.com/products/boy-brow" }
  ]
}

Agents can also send productUrls as a plain string list.

Field	Default	Meaning
detectChanges	true	Compare against the last snapshot
snapshotStoreName	jsonld-product-snapshots	Named store for scheduled runs
respectRobotsTxt	true	Skip paths the site disallows
maxConcurrency	3	Keep this low

Output

One dataset row per extracted Product offer.

{
  "url": "https://www.glossier.com/products/boy-brow",
  "status": "ok",
  "name": "Boy Brow",
  "price": 22,
  "currency": "USD",
  "availability": "InStock",
  "changeType": "price_down",
  "previousPrice": 24,
  "error": null
}

status values:

• ok — Product JSON-LD with a usable price or availability
• incomplete_offer — Product JSON-LD without price and availability
• no_product_schema — page has no Product JSON-LD (common on category/configurator pages)
• http_error — non-2xx or network failure
• robots_blocked — path disallowed by robots.txt
• parse_error — HTML fetched but JSON-LD could not be parsed

changeType values: new, unchanged, price_up, price_down, back_in_stock, out_of_stock, offer_changed.

What this is not

• Not an Amazon, Walmart, eBay, or Google Shopping crawler
• Not a search scraper. You must already know the product URL
• Not a login or cart tool
• Not a CSS/Playwright fallback. If the shop hides offer data from JSON-LD, the row is no_product_schema or incomplete_offer

That honesty is the product. Competing “any store” Actors often claim Amazon coverage they cannot keep reliable.

Pricing

Pay per event:

Event	What it bills	Suggested price
page_checked	Every URL fetched or blocked after work	$0.001
product_record	Every structured product row written	$0.002
change_detected	Every row that is not unchanged	$0.003
apify-actor-start	Platform start (keep default)	$0.00005

A 100-URL daily monitor with no changes costs about $0.30 in events (100 × $0.001 + 100 × $0.002). A day with 10 real changes adds about $0.03.

Competitor dump parsers charge $2–$5 per 1,000 rows and do not emit change events. This Actor is cheaper on a check and charges extra only when something moved.

Platform cost for Cheerio HTTP is typically well under $0.0005 per URL. Event prices above that leave a positive margin after Apify’s 20% store cut.

Schedule

1. Save your URL list as the Actor input
2. Keep snapshotStoreName the same
3. Run daily or hourly from Apify Schedules
4. Filter the dataset on changeType != unchanged

Limits

• Public HTML only
• JSON-LD Product / Offer / AggregateOffer only
• Some shops put Product schema on category or builder pages without a real offer
• Sites that block anonymous HTML fetch will return http_error
• robots.txt is respected when enabled

Agent / MCP

Required input is only URLs. Output is a flat typed row. No browser, no cookies, no extra steps.