Wix Marketplace Professionals Metadata Scraper

Extracts public listing-card metadata from Wix Marketplace's "Hire a Professional" pages (https://ja.wix.com/marketplace/hire). It reads only what is already visible on the public listing pages — no login, no detail-page visits, no contact data.

What this Actor does

• Reads public Wix Marketplace listing pages under https://ja.wix.com/marketplace/hire (and its ?page=N pagination).
• Parses provider/listing cards using Cheerio (static HTML parsing only).
• Extracts listing-card metadata: name, partner ID, a constructed profile URL (when safely derivable), location, rating, review count, starting price, short description, and profile image URL.
• Writes results to the default dataset and a RUN_SUMMARY object to the default key-value store on every run.

What this Actor does NOT do

• This Actor is not a contact scraper. It does not extract emails, phone numbers, contact form URLs, social links, or external website URLs.
• It does not visit provider detail pages in v0 — only the listing pages themselves are read.
• It does not use login, cookies, sessions, a proxy, browser automation (no Playwright/Puppeteer), or any CAPTCHA bypass.
• It does not crawl any site other than the Wix Marketplace listing pages described above.
• It does not download profile images — only the image URL, if present, is recorded.

Input

Field	Type	Default	Description
maxItems	integer	20	Maximum number of listing records to extract (1–100).
startPage	integer	1	First Wix Marketplace listing page to read (1–42).
maxPages	integer	3	Maximum number of listing pages to read (1–42).
debug	boolean	false	Enables verbose debug logging.

Output fields

Each dataset item contains:

• professionalName
• partnerId
• profileUrl — constructed only when a partner ID or profile path can be safely derived from the listing page source; otherwise null.
• location
• rating
• reviewCount
• startingPrice
• shortDescription
• profileImageUrl
• sourceUrl
• page
• found
• retrievedAt

Missing optional fields are set to null. Values are never invented.

Output limits and cost control

To protect against runaway usage and negative-profit runs on Pay-Per-Result pricing, the Actor computes:

effectiveMaxItems = min(input.maxItems, ACTOR_MAX_PAID_DATASET_ITEMS, 100)

• ACTOR_MAX_PAID_DATASET_ITEMS is read from the platform environment when present; otherwise it is ignored.
• 100 is an internal hard cap that always applies.
• The Actor stops pushing dataset records once effectiveMaxItems is reached, and stops requesting further listing pages once the cap is hit — it never crawls pages unnecessarily after the output limit has been reached.
• startPage / maxPages are always respected, and the Actor never reads past listing page 42.

RUN_SUMMARY v1

Every run writes a RUN_SUMMARY object (key: RUN_SUMMARY) to the default key-value store with the following schema:

{
  "version": "RUN_SUMMARY v1",
  "totalItems": number,
  "requestedMaxItems": number,
  "effectiveMaxItems": number,
  "actorMaxPaidDatasetItems": number | null,
  "skippedDueToLimit": number,
  "errorCount": number,
  "warningCount": number,
  "warnings": string[],
  "sourceUrl": string,
  "startPage": number,
  "maxPages": number,
  "lastPageAttempted": number,
  "requestsFinished": number,
  "requestsFailed": number,
  "cardsSeen": number,
  "recordsMatched": number,
  "recordsSkipped": number,
  "retrievedAt": string
}

Notes on severity

• Reaching maxItems (i.e. effectiveMaxItems) is not an error — it is reflected in skippedDueToLimit.
• Fewer results than requested is not automatically an error; the source listing may simply have fewer matching cards.
• A missing professionalName causes that card to be skipped (counted in recordsSkipped) rather than failing the run.
• If a profileUrl cannot be safely constructed, the record is kept with profileUrl: null and a single run-level warning is added (not one warning per affected record).

Limitations

• Relies on the current public HTML structure of the Wix Marketplace listing pages (data-hook="provider-container", provider-box-<uuid>, partner-business-name-<uuid>, and similar patterns). If Wix changes this structure, parsing may need updates.
• Only listing-card data is extracted; no per-provider detail-page enrichment is performed in v0.
• Some optional fields (location, rating, review count, starting price, short description, profile image) may be null if not present or not reliably parseable on a given card.

Development status

v0 — initial implementation. Static Cheerio-based parsing only. Not yet published.