Ebay Seller Scraper

What does eBay Seller Scraper do?

This Actor extracts seller information and visible listing cards from public eBay seller or seller-search pages. It accepts official HTTPS /usr/, /str/, and /sch/ URLs, or seller usernames that are converted to public /usr/ profile URLs. The output separates seller profiles from seller listings so downstream workflows can associate feedback identity with the public items shown on the page.

The Actor is a public-page extractor, not an account or access-control bypass. It does not log in, submit feedback actions, solve CAPTCHA or device challenges, defeat WAF controls, use private endpoints, or save raw HTML. If eBay returns a public access boundary, run-level counts and logs record it; the challenge page is never added to the dataset.

Output

Normal seller profile rows use recordType: "seller-profile" and can include sellerUsername, sellerName, feedbackScore, positiveFeedbackPercent, and sellerProfileUrl. Normal listing rows use recordType: "ebay-seller-listing" and can include sellerUsername, listingTitle, listingId, listingUrl, price, currency, condition, shippingText, category, and imageUrl. Both row types include stable recordId, sourceUrl, url, scrapedAt, httpStatus, accessStatus, ok, found, dataAvailable, usedProxy, and sourceTransport provenance.

Optional fields are omitted when the public page does not expose them. Unknown values are not turned into zero, false, or placeholder identity. Seller and listing IDs are deduplicated before maxItems is applied.

You can download the dataset in various formats such as JSON, HTML, CSV, or Excel. The dataset contains source-backed seller and listing rows only. OUTPUT_SUMMARY in the key-value store reports status and runStatus (SUCCEEDED or FAILED) and dataStatus (COMPLETE, PARTIAL, FAILED, or NO_RESULTS), saved dataset count, diagnostic count, dataset write failures, unique record count, request totals, proxy use, and access-boundary detection.

Input

Field	Type	Default	Description
startUrls	array	—	Up to 50 official HTTPS /usr/, /str/, or /sch/ URLs.
sellerUrls	array	—	Backward-compatible alias for seller/search URLs.
sellerUsernames	array	—	Up to 50 public seller usernames.
maxItems	integer	50	Normal seller/listing rows to emit, from 1 to 500.
timeoutSecs	integer	30	Request timeout from 5 to 120 seconds.
maxBytes	integer	5000000	Response limit from 100,000 to 15,000,000 bytes.
maxRetries	integer	1	Bounded transient retry count from 0 to 3.
userAgent	string	standard browser	Optional public request User-Agent.
proxyConfiguration	object	—	Optional permitted Apify proxy configuration.

At least one of startUrls, sellerUrls, or sellerUsernames is required. Inputs are validated strictly: non-HTTPS URLs, non-eBay hosts, item URLs, embedded credentials, fragments, unknown fields, and coerced numeric values are rejected.

How to run it

Start with one public seller profile URL and a modest maxItems value. For example, use https://www.ebay.com/usr/harbor-tools when that profile is publicly reachable, or provide sellerUsernames: ["harbor-tools"]. Inspect recordType before loading rows into a database. Use diagnosticCount, accessBoundaryDetected, and the run logs to understand blocked, empty, or failed sources; diagnostic rows are never written to the dataset.

Reliability, cost, and ethical use

Each request has a bounded timeout, response-size limit, redirect limit, and transient retry limit. A proxy may change routing but never changes the fail-closed boundary behavior. Runs are charged according to normal Apify compute and transfer usage; fewer URLs, smaller limits, and focused profiles generally finish sooner. Respect eBay terms, robots instructions, regional restrictions, and applicable privacy law. Public seller names, feedback, and locations can still be personal or commercially sensitive data. Use the results only for a legitimate purpose and never attempt to infer private attributes or circumvent access controls.