eBay Sold Comps Scraper

Collect publicly visible sold and ended eBay listings as pricing comparables. Supply completed-search URLs or known ended item URLs; the Actor normalizes each result into one record with the item number, title, final price and display text, currency, condition, sold or ended date, sale type, seller, and image. Search URLs are always forced to LH_Sold=1 and LH_Complete=1, requests are bounded by timeout, response size, retry count, and page count, and a sign-in wall or challenge is reported as an access boundary instead of being bypassed.

Get started

1. Open the Actor Store page and select Try for free or open the Input tab.
2. Paste one or more https://.../itm/... item URLs into startUrls, or leave startUrls empty and rely on searchUrl for a completed-search crawl.
3. Keep maxItems and maxPages small for the first run. The saved run used seven item URLs with maxItems: 5 and maxPages: 2.
4. Start the Actor, open the dataset, and export the comparables as JSON, CSV, or Excel. Open Run summary (OUTPUT_SUMMARY) to check counts, retries, and access-boundary state.

Input

All fields are optional, so an empty input still runs the default iPhone completed search. Unsupported field names and out-of-range values are rejected before requests start.

Field	Type	Required?	Default	Meaning
startUrls	object array	Optional	Not supplied	Up to 20 { "url": "..." } entries. Each must be an HTTPS eBay item (/itm/) or search (/sch/, /s/) URL; search entries get LH_Sold=1 and LH_Complete=1 added. When empty, searchUrl is used.
searchUrl	string	Optional	https://www.ebay.com/sch/i.html?_nkw=iphone&_sacat=0&LH_Sold=1&LH_Complete=1	Completed-search URL used when no startUrls are supplied. The sold and completed filters are always forced on, up to 2,000 characters.
maxItems	integer	Optional	25	Maximum comparable records saved across all sources, 1–100.
maxPages	integer	Optional	3	Maximum search pages followed per source, 1–5; item URLs use one page.
timeoutSecs	integer	Optional	20	Timeout per bounded request or browser navigation, 5–120 seconds.
maxBytes	integer	Optional	5000000	Largest response body accepted, 100,000–10,000,000 bytes.
maxRetries	integer	Optional	1	Extra attempts for timeouts and retryable responses (408, 425, 429, 5xx), 0–3. Challenges and access denials are never retried as transient errors.
browserFallback	boolean	Optional	true	Render the same public URL in an anonymous browser only after a successful, unchallenged HTTP response contained no comparable evidence.
browserWaitMs	integer	Optional	2000	Wait after DOMContentLoaded before reading rendered evidence, 0–10,000 ms.
userAgent	string	Optional	Built-in desktop Chrome string	User-Agent sent with direct requests, up to 300 characters.
proxyConfiguration	object	Optional	Not supplied	Standard Apify Proxy configuration. With a proxy configured, a challenged page is retried on fresh sessions, with at most two attempts for a sign-in wall, after which the boundary is reported.

Example input, matching the sample stored in .actor/input.json:

{
  "startUrls": [
    {
      "url": "https://www.ebay.co.uk/itm/377440218011"
    }
  ],
  "maxItems": 5,
  "maxPages": 1,
  "timeoutSecs": 30,
  "browserFallback": false
}

Omitted fields fall back to the schema defaults in the table, so this sample deliberately narrows the run to five comparables on one page and turns the browser fallback off. Item URLs are read as ended listings; only the URLs you supply are visited, and maxItems also caps how many seeded URLs are visited. The saved run added maxRetries: 1 and a proxy configuration, and set maxPages: 2.

Output

The dataset contains one row per unique comparable, de-duplicated by canonical item URL. A row is saved only when the page or card provides an item number, a title, and a parseable final price; pages that expose nothing usable produce a diagnostic in the run summary, not a dataset row.

Field	Type	Availability	Meaning
recordType	string	Always	Record label; always sold-comparable.
recordId	string	Always	Stable identifier, ebay-sold-comparable-<itemId>, or a short hash when no item number is available.
sourceUrl	string	Always	Exact page that produced the row: the item URL or the completed-search page.
httpStatus	integer	Always	HTTP status of the successful response, normally 200.
accessStatus	string	Always	Access outcome; only AVAILABLE rows are saved.
ok	boolean	Always	Request-level success flag, always true for a saved row.
found	boolean	Always	Whether comparable evidence was found, always true for a saved row.
dataAvailable	boolean	Always	Whether source-backed data was saved, always true.
usedProxy	boolean	Always	Whether the response came through a configured proxy.
sourceTransport	string	Always	How the page was read: direct, apify-proxy, or browser.
url	string	Always	Canonical https://<host>/itm/<itemId> listing URL.
itemId	string	Always	eBay item number parsed from the listing URL.
title	string	Always	Listing title as displayed on the item page or search card.
price	number	When published	Final listing price as a number, with separators removed and decimals normalized.
priceText	string	When published	Price exactly as displayed, for example US $89.30, preserving the marketplace's currency notation.
currency	string	When published	Three-letter currency code taken from the price text, or inferred from the $, £, or € symbol.
condition	string	When published	Condition label shown on the page, passed through as displayed.
soldDate	string	When published	Sale or end date in eBay's short form, for example Wed, 16 Sep at 7:49 AM, without a year or time zone.
saleType	string	When published	sold for a confirmed sale, ended for a listing that closed without a recorded sale.
shippingText	string	When published	Shipping or returns wording from a search card; item pages do not publish it in this position.
sellerName	string	When published	Seller display name when the item page or card exposes it.
imageUrl	string	When published	Primary image URL on an ebayimg.com or ebaystatic.com host.
scrapedAt	string	Always	Time the row was saved, in ISO UTC.

The example below is a real saved row from the 2026-10-25 run, complete as stored.

{
  "recordType": "sold-comparable",
  "recordId": "ebay-sold-comparable-820007572989",
  "sourceUrl": "https://www.ebay.co.uk/itm/820007572989",
  "httpStatus": 200,
  "accessStatus": "AVAILABLE",
  "ok": true,
  "found": true,
  "dataAvailable": true,
  "usedProxy": true,
  "sourceTransport": "apify-proxy",
  "url": "https://www.ebay.co.uk/itm/820007572989",
  "itemId": "820007572989",
  "title": "8.8 Differential Cover Rear End Girdle System for Ford Mustang F-150",
  "price": 89.3,
  "priceText": "US $89.30",
  "currency": "USD",
  "condition": "New",
  "soldDate": "Wed, 16 Sep at 7:49 AM",
  "saleType": "sold",
  "sellerName": "Trendsphereny",
  "imageUrl": "https://i.ebayimg.com/images/g/3VsAAeSwW4tqgQX-/s-l400.jpg",
  "scrapedAt": "2026-10-25T20:43:12.190Z"
}

Notes from the other four rows of the same run: the first row (227421081900, GBP 21.25) has no sellerName or shippingText; 377440218011 is a GBP 7,000.00 vehicle recorded as saleType: "ended" with the seller rebre4680; and 407137704260 retains condition: "--" because that is the text eBay displayed for the condition slot. shippingText appears in the schema but no saved item page exposed it, so comparables taken from item URLs may never carry shipping details. soldDate is a display string rather than an ISO timestamp, so parse it with the marketplace locale in mind.

Run summary

Run summary is stored separately from the dataset in the run's key-value store under OUTPUT_SUMMARY.

Field	Meaning
status	Actor process outcome: SUCCEEDED or FAILED.
runStatus	Canonical run status; matches status.
dataStatus	Completeness of saved rows: COMPLETE, PARTIAL, NO_RESULTS, or FAILED.
datasetItems	Rows actually saved.
diagnosticItems	Diagnostic dataset rows; always 0.
diagnosticCount	Number of operational issues recorded.
errorCount	Same count expressed as errors; equals diagnosticCount.
diagnostics	Array of { url, error, errorCode, scrapedAt } entries kept outside the comparable dataset.
uniqueItems	Distinct recordId values among saved rows.
sourceCount	Number of seeds used, meaning startUrls entries or 1 when only searchUrl was used.
requestCount	Requests attempted, including retries and browser navigations.
successfulRequests	Requests that returned a usable 2xx response without an access boundary.
failedRequests	Failed attempts, including retries that later succeeded.
proxyUsed	Whether a proxy configuration was active for the run.
accessBoundaryDetected	Whether a login wall, CAPTCHA, rate limit, geofence, or unsafe redirect was seen.
pagesVisited	Pages processed, including pages that later failed.
completedAt	Time the summary was written, in ISO UTC.

Status	Data status	Meaning
SUCCEEDED	COMPLETE	Comparables were saved and no diagnostic was recorded.
SUCCEEDED	PARTIAL	At least one comparable was saved, but another page or record failed. Saved rows remain available.
SUCCEEDED	NO_RESULTS	Every source was read cleanly but produced no sold or ended comparable.
FAILED	FAILED	A fatal input error, or every source failed before a comparable was saved.

Example OUTPUT_SUMMARY from the 2026-10-25 run:

{
  "status": "SUCCEEDED",
  "runStatus": "SUCCEEDED",
  "dataStatus": "COMPLETE",
  "datasetItems": 5,
  "diagnosticItems": 0,
  "diagnosticCount": 0,
  "errorCount": 0,
  "diagnostics": [],
  "uniqueItems": 5,
  "sourceCount": 7,
  "requestCount": 7,
  "successfulRequests": 5,
  "failedRequests": 2,
  "proxyUsed": true,
  "accessBoundaryDetected": false,
  "pagesVisited": 5,
  "completedAt": "2026-10-25T20:43:34.935Z"
}

In this run requestCount (7) exceeds pagesVisited (5) and failedRequests is 2, because two item requests needed a retry before returning a usable page. No diagnostic was recorded, so the run still reports COMPLETE: a retried request that eventually succeeds is not a data gap.

Coverage and limitations

The Actor reads only the item and completed-search pages you provide, up to maxPages search pages per seed and maxItems records overall. eBay's sold and completed filter is frequently answered with a sign-in wall for shared datacenter addresses; when that happens the Actor reports LOGIN_REQUIRED in diagnostics, sets accessBoundaryDetected, and stops that source. It does not sign in, solve CAPTCHAs, or invent a comparable.

saleType reflects the page wording: a confirmed sale is sold, while a listing that was withdrawn or closed without a sale is ended. Both are useful price evidence but they are not the same event, so filter on saleType before averaging. Search-card rows can include shippingText but no saleType; item-page rows can include saleType and sellerName but usually no shipping wording. soldDate is eBay's short display text without a year, and condition is whatever label the page shows, including placeholder dashes. The optional anonymous browser fallback runs only for a successful, unchallenged page that yielded no evidence, and its cost is much higher than a direct request.

Usage and support

Direct HTTP requests are the cheap path; enable browserFallback only when item pages render their comparable details with JavaScript. Cost grows with seeds, pages, retries, and any proxy or browser usage, so start with a handful of item URLs, inspect the dataset and OUTPUT_SUMMARY, then scale. See the Actor's Pricing tab for current account rates.

Use the Actor Issues page to report a problem. Include the run ID, the input used, and the OUTPUT_SUMMARY; do not include tokens, cookies, or private data. This Actor is not affiliated with, endorsed by, or sponsored by eBay. Use it only for public data and in accordance with eBay's terms, robots guidance, and applicable law. Sold listings can reveal seller and buyer activity; use the output only for lawful market-research purposes.