eBay Marketplace Scraper: Listings, Prices, Sellers & Item Details

eBay Marketplace Scraper turns eBay into structured listing data across sixteen country storefronts. Search by keyword, seller username or category id, or paste links you already have: result pages, seller listing pages, browse category pages and individual item pages are all accepted and can be mixed in one run. Every listing comes back as one flat row with the price, the buying format, delivery, the seller and their feedback score, and, when you ask for it, the full image gallery, the item specifics table and the seller's own description. Export to JSON, CSV or Excel, or pull the results straight into your app through the API.

Why This Scraper?

• Sixteen marketplaces, each checked live. United States, United Kingdom, Germany, Australia, Canada, France, Italy, Spain, Ireland, Austria, Switzerland, Netherlands, Poland, Hong Kong, Singapore and Malaysia. Every one was fetched and confirmed to return real, id-bearing listings before it was offered as an option, and prices come back in that storefront's own currency. (The Philippines is not offered: ebay.ph redirects to the United States storefront, so listing it separately would be an option that quietly does something else.)
• Around 45 fields per listing, including the low and high end of a variant price range, the pre-discount price and computed discount percentage, bid count and time left on auctions, the sold count, the item location, and the seller's positive-feedback percentage and score. Most of that arrives from the result page itself, so it costs nothing extra.
• Four ways in. Keyword search, a whole seller's store, a category id, or pasted links. Search filters use the marketplace's own native filters (condition, buying format, price band, delivery, returns, item location, sort order) so they narrow the result set at the source rather than throwing rows away after you have paid for them.
• A search that matches nothing returns nothing. eBay answers a no-match query with a full page of loosely related suggestions that sit in the same result list as genuine matches. This scraper reads the page's own result count and returns zero rows instead of handing you sixty listings that have nothing to do with your query.
• Incremental mode for recurring runs. Schedule the same search daily and get only what is NEW or UPDATED by default, with the changed field names listed on every updated row; turning on Emit expired also gets you EXPIRED and, on a later run, REAPPEARED. Auction countdowns and walk bookkeeping never count as a change, so a ticking clock or a re-ordered page never bills you for a change that did not happen.
• Resume a large pull from a previous run id without re-returning or re-charging for listings you already have.

Use Cases

• Resellers and arbitrage sellers: track prices, discounts and sold counts across marketplaces to spot underpriced inventory worth flipping.
• Price monitoring tools: feed a comparison site or a price-tracking bot with live listings converted to each storefront's own currency.
• Sourcing and catalog research: pull a whole seller's store with item specifics and full descriptions before deciding what to list elsewhere.
• Listing and brand monitoring: watch a category or keyword for new and changed listings on a recurring schedule.
• Market and demand research: track sold counts and auction activity for a keyword over time with Incremental mode.

Data You Get

Sample shape, values are illustrative placeholders, not from a live listing.

Field	Example
itemId	"000000000001"
url	"https://www.ebay.com/itm/000000000001"
title	"Sample Listing Title"
condition	"Used"
brand	"Sample Brand"
marketplace	"com"
price	100.0
currency	"USD"
listingType	"buy_it_now" (also "auction", "best_offer")
shippingText / freeShipping	"Free delivery" / true
itemLocation	"United States"
returnsText / freeReturns	"Free returns" / true
soldCount	42
sellerName	"examplestore"
sellerFeedbackPercent / sellerFeedbackCount	98.5 / 1234
badges	["Sample Badge"]
imageUrl	"https://i.ebayimg.com/images/g/000000000000/s-l500.webp"
attributes	["Buy It Now", "Free delivery", "Located in United States", "Free returns", "42 sold"]
itemSpecifics	{"Processor": "Sample Processor", "SSD Capacity": "1 TB"} (when Read item pages is on)
description	"Full seller description text appears here when Read item pages is on."
detailIncomplete	true (only when an item page, or just its separate description document, could not be read after a retry; see "Reading item pages and the surcharge" below)
changeType	"NEW" (also "UPDATED", "UNCHANGED", "REAPPEARED", "EXPIRED"; Incremental mode only)

Every listing also carries subtitle (the seller's own strapline, kept separate from the condition row), conditionId (eBay's numeric condition code, populated only when the condition wording is in English), marketplaceDomain, priceMax and priceText for variant price ranges, originalPrice and discountPercent when the card shows a strikethrough price, bidCount and timeLeft on auctions, shippingCost alongside shippingText, sellerUrl, searchQuery, sourceUrl and pageNumber for provenance, and scrapedAt for the run's own timestamp. Reading item pages additionally fills images (the full gallery), categoryPath, mpn, model, color, gtin, availability, quantityAvailable, quantitySold, paymentMethods, shippingDetails, returnPolicy, listPrice, listedAt, updatedAt and sellerNotes; all of these stay null unless Read item pages is on. Incremental mode adds changedFields (the field names that changed on an UPDATED row), firstSeenAt and lastSeenAt.

How to Use

1. Pick a mode: search (by keyword, seller username, category id and native filters) or url (paste links).
2. Fill in the fields for that mode, add any filters and a sort order, and turn on Read item pages if you want the image gallery, item specifics and full description.
3. Set Max listings to control run size and cost (0 for unlimited), then click Start.
4. Download the dataset as JSON, CSV or Excel, or read it through the API.

Search one marketplace by keyword:

{
  "mode": "search",
  "marketplace": "com",
  "queries": ["mechanical keyboard"],
  "maxItems": 50
}

Search with the marketplace's own filters, including a multi-select condition:

{
  "mode": "search",
  "marketplace": "co.uk",
  "queries": ["vinyl record"],
  "condition": ["new", "used"],
  "buyingFormat": "buy_it_now",
  "minPrice": 10,
  "maxPrice": 60,
  "freeShippingOnly": true,
  "sortBy": "newly_listed",
  "minSellerFeedbackPercent": 98,
  "maxItems": 100
}

Pull a whole seller store with item pages:

{
  "mode": "search",
  "marketplace": "de",
  "sellerUsernames": ["examplestore"],
  "fetchDetails": true,
  "maxItems": 200
}

Paste links, mixing marketplaces and link kinds:

{
  "mode": "url",
  "urls": [
    "https://www.ebay.com/sch/i.html?_nkw=laptop&LH_BIN=1",
    "https://www.ebay.co.uk/sch/examplestore/m.html",
    "https://www.ebay.de/itm/000000000001",
    "https://www.ebay.com/b/PC-Laptops-Netbooks/177",
    "https://www.ebay.com/sch/177/i.html"
  ],
  "maxItems": 150
}

Supported links: result pages, category-scoped searches (/sch/<slug>/<categoryId>/i.html), /sch/<categoryId>/i.html, seller pages, /str/ and /usr/ seller store links (scraped as that seller's listings), /p/<id> catalog product pages (scraped as that product's listings), /b/ browse pages and item links in any of the usual forms. The urls field is pre-filled with example links covering every shape (item variants such as a title slug, a tracking query or an m. host point at the same listing, and category-id or seller-path links redirect to a page already listed, so they are scraped once). If you switch to URL mode and leave it untouched, the distinct examples (search, country host, paged and filtered search, browse category, seller, item) are scraped, so replace it with your own links.

Run it from your code

Python:

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("abotapi/ebay-com-scraper").call(
    run_input={"mode": "search", "queries": ["mechanical keyboard"], "maxItems": 20})
for listing in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(listing["title"], listing["price"], listing["currency"])

JavaScript:

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: '<YOUR_APIFY_TOKEN>' });
const run = await client.actor('abotapi/ebay-com-scraper').call({
  mode: 'search', queries: ['mechanical keyboard'], maxItems: 20,
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();

Or connect it to Make, Zapier, n8n, Google Sheets or webhooks from the Integrations tab.

Reading item pages and the surcharge

Turning on Read item pages adds one extra request per listing and bills one Item page event for every listing that is both pushed to the dataset and had its own item page successfully read. A listing whose item page is refused, and stays refused after one retry pass, ships with its result-card fields only, carries detailIncomplete: true, and is not charged the surcharge. When the item page itself is read but only its separate description document keeps failing, the listing keeps its item-specifics, images and category trail, is still stamped detailIncomplete: true for the missing description, and is still charged: the surcharge follows whether the item page was read, not whether every last field on it came back. A pasted item link has no result card behind it, so its item page is always read regardless of the Read item pages toggle, and is billed the Item page event whenever that read succeeds, even with the toggle off; it follows the same refused-item-page and description-only rules above. An UNCHANGED row suppressed by Incremental mode is never billed the surcharge, even on a run where its item page had to be read to compute the fingerprint. Seller feedback comments (reviews, see Max seller feedback comments per listing) are off by default and ride on the same event when you turn them on: the seller's feedback page is read once per seller per run, and each started block of 10 delivered comments adds one more Item page event (up to 5 events for 50 comments), billed once per seller on the first of that seller's listings that is pushed with its item page read. Later listings from the same seller carry the same comments and add nothing. A listing with no readable feedback, a suppressed UNCHANGED row, or maxReviewsPerProduct left at 0 adds nothing. reviews is not part of change detection in Incremental mode, because the newest comments move constantly.

Max listings, Max pages and page size

Max listings is the whole run's cap; Max pages limits how many result pages are read per keyword, seller or link before moving on, independent of the cap. When the whole run fits inside one small result page, the actor asks eBay for a 60-item page instead of a 240-item one, since a 240-item page is around 4.7 MB of transfer to keep 20 rows. Raising Max listings past 60 switches it back to the large page automatically.

Resume and recurring updates

• Resume (resumeFromRunId) continues one interrupted run: paste its run or dataset ID and the actor skips every listing id it already returned there, so you don't pay twice. It cannot be combined with Incremental mode once that mode already has saved state for the same search: the run fails with a clear message rather than guessing which one you meant.
• Incremental mode (incrementalMode) is for scheduled runs over the same search. The first run returns every matching listing as NEW. Later runs classify each listing NEW, UPDATED (with changedFields), or UNCHANGED (suppressed and not billed unless emitUnchanged is on). EXPIRED rows are only produced once a run has scanned the tracked search to its natural end, with no cap, no resume and at least one listing seen, and only with emitExpired on; a capped, resumed or empty-scan run keeps the existing state untouched instead of guessing. REAPPEARED is never a default outcome of turning Incremental mode on: it only fires for a listing that was previously tombstoned EXPIRED by that same complete-scan process and is then seen again on a later run. stateKey names or shares one monitoring campaign; left empty, the actor derives one from the mode, marketplace, keywords, sellers, links and filters, so different searches never collide. Incremental mode also excludes timeLeft, attributes, scrapedAt, updatedAt, sourceUrl, pageNumber and searchQuery from change detection, because an auction countdown or a walk artifact would otherwise mark a listing changed on every single run; every real change still surfaces through a tracked structured field such as listingType, bidCount or shippingCost. Add your own noisy fields to Ignore fields when detecting changes (ignoreFieldsForChanges, for example soldCount): a plain name is ignored at any depth, a dotted path such as shippingDetails.cost targets one nested field. Ignored fields stay in the output but never trigger UPDATED or appear in changedFields. The comparison re-reads the stored listing with the current ignore list, so editing it causes no one-time wave of UPDATED rows. A pasted link that starts after page 1 (_pgn above 1) never produces EXPIRED rows, because the earlier pages were not scanned.

Send results into your apps (MCP connectors)

Results can also be piped straight into the apps you already use, over Model Context Protocol. Authorize a connector under Apify, Settings, API & Integrations, then select it in the Export to your apps (MCP connectors, optional) section of the input.

• mcpConnectors picks one or more authorized connectors. Leave it empty and nothing happens: the export code never runs and the dataset is byte for byte identical.
• notionParentPageUrl is required for the Notion connector, which gets one page per listing, and is ignored by every other connector.
• maxNotifyListings caps how many listings are written to each connector per run. It does not affect the dataset.

What gets written is a condensed, human readable summary per listing: a title plus the key fields flattened to plain text. Nested objects collapse to their main value and arrays are trimmed to a few entries. The complete record always stays in the Apify dataset.

If a connector write fails, the run logs a warning and carries on. Scrape data is never affected.

Input Parameters

Parameter	Type	Default	Description
mode	string	search	search or url.
queries	array	no default (prefill: ["laptop"])	Search keywords. Search mode only.
sellerUsernames	array	(none)	Seller usernames whose listing pages are walked. Search mode only.
marketplace	string	com	Which of the sixteen storefronts to search. Search mode only; a pasted link carries its own.
categoryId	string	(none)	Numeric category id from the marketplace's own category links. A category name is not accepted: it is ignored with a warning in the run log and the search runs without a category filter.
condition	array	(none)	New, Open box, the four refurbished grades, Seller refurbished, Used, For parts. Multi-select.
buyingFormat	string	any	Any, Buy It Now, Auction, Accepts offers.
minPrice	integer	(none)	Keep only listings priced at or above this amount, in the marketplace's own currency.
maxPrice	integer	(none)	Keep only listings priced at or below this amount, in the marketplace's own currency.
freeShippingOnly	boolean	false	Keep only listings flagged as free delivery.
freeReturnsOnly	boolean	false	Keep only listings offering free returns.
returnsAcceptedOnly	boolean	false	Keep only listings that accept returns at all.
itemLocation	string	default	Marketplace default, marketplace country only, same continent, or worldwide.
sortBy	string	best_match	Best match, ending soonest, newly listed, price and delivery lowest or highest, or distance nearest (with a postal code).
sortOrder	string	none	Alias of sortBy (timeNewlyListed, pricePlusPostageLowest, pricePlusPostageHighest, distanceNearest).
postalCode	string	none	Origin for the distance-nearest sort.
keywords	array	none	Alias of queries, used when queries is empty.
ebaySite	string	none	Alias of marketplace, for example ebay.co.uk.
subcategoryId	string	none	Numeric category id that overrides categoryId.
itemCondition	string	any	any, new or used, used when condition is empty.
conditionId	integer	none	Marketplace condition id, overrides itemCondition.
aspectFilter	object	none	Item-specific facets such as {"Brand": "Nike"}.

| urls | array | one example of each supported link shape | Result, seller, browse category (/b/... and /sch/<categoryId>/i.html) and item links (bare id, title slug, tracking query, m. host, country hosts). URL mode only. Leaving the field untouched scrapes the distinct example links (search, country host, paged search, browse category, seller, item), so clear it or paste your own. /str/ and /usr/ links are scraped as the seller's listings. /p/<id> catalog links are scraped as that product's listings. | | minSellerFeedbackPercent | integer | (none) | Applies in both modes. Sellers with no score yet are kept. | | fetchDetails | boolean | false | Read each listing's item page. Adds one Item page event per listing whose item page is actually read. Refused item pages get a second pass; a listing that still cannot be read ships without details, flagged detailIncomplete, unbilled. | | maxReviewsPerProduct | integer | 0 | With item pages read, attach up to this many of the seller's most recent public feedback comments to each listing as reviews (reviewer, rating, date, text, item, price; helpful is always null because the site has no helpful votes). 0 (the default) turns it off, maximum 50. The seller's feedback page is read once per seller per run; every started block of 10 delivered comments adds one Item page event, billed once per seller on that seller's first returned listing, and later listings from the same seller reuse the comments at no extra charge. | | maxItems | integer | 20 | The run's cap. 0 means unlimited. | | maxPages | integer | 0 | Result pages per keyword, seller or link. 0 means no limit of your own: the run still stops at Max listings, at the end of results, or at a built-in ceiling of 100 pages. | | resumeFromRunId | string | (none) | Continue one interrupted run from its run or dataset id. | | incrementalMode | boolean | false | Return only what changed since the last run of the same search. | | ignoreFieldsForChanges | array | none | Field names or dotted paths excluded from change detection (incremental mode only). | | stateKey | string | (none) | Name a monitoring campaign, or deliberately share state. | | emitUnchanged | boolean | false | Also return, and bill, UNCHANGED rows. | | emitExpired | boolean | false | Also return, and bill, EXPIRED rows. | | proxy | object | Apify Proxy, residential | Connection configuration. Residential is the default; free plans continue on a backup connection. | | mcpConnectors | array | (none) | Optional MCP connectors to also write results into. | | notionParentPageUrl | string | (none) | Notion parent page id or URL, Notion connector only. | | maxNotifyListings | integer | 50 | Cap on items written to each connector per run. |

Sold and completed listings. The marketplace shows sold and completed listings only to signed-in visitors: anonymous requests for them are answered with an interruption page even on a warm residential session (checked 2026-10-19). This actor does not sign in, so it returns active listings only and has no sold price or end date fields. API callers migrating from other actors can still send daysToScrape, includeCompletedListings, sortOrder: endedRecently, buyingFormat values all, buyItNow and acceptsOffers, and itemLocation: domestic: they are accepted in code (the sold-history ones log a warning and return active listings), but they are not offered in the Console form.

Additional output fields. Each row also carries keyword, itemNumber, buyingFormat, bids, thumbnailUrl, fullResThumbnailUrl, sellerUsername, sellerFeedbackScore, sellerPositivePercent, shippingPrice, shippingCurrency, shippingType, totalPrice (price plus shipping), category, categoryId and breadcrumbs, productDetails, descriptionIframeSrc and sellerDescriptionText. Most repeat values the row already holds under these names; sellerUsername is the seller handle (read from the item page's feedback link or the result card, never the display name), while categoryId and descriptionIframeSrc are new data read from the item page (empty when item pages are not read). All of them are empty when the source has no value. They are excluded from change detection in monitoring mode, as is reviews; the tracked fields quantityAvailable, quantitySold, sellerName, sellerFeedbackPercent and sellerFeedbackCount are compared, and a first appearance of one of them on a listing that held no value before never counts as a change.

Views. The dataset has four views: Overview (the listing card), Item details (gallery, specifics, seller card, quantity available and sold), Changes (Incremental mode state, changedFields) and Seller feedback (reviews). Select one with the view parameter in the dataset API or the tabs in the Console.

Output Example

Sample shape, values are illustrative placeholders, not from a live listing. Shown with Read item pages and Incremental mode both on so every field appears; a default run omits changeType, changedFields, firstSeenAt and lastSeenAt, and the item-page-only fields stay null.

{
  "itemId": "000000000001",
  "url": "https://www.ebay.com/itm/000000000001",
  "title": "Sample Listing Title",
  "subtitle": "Sample seller strapline",
  "condition": "Used",
  "conditionId": "3000",
  "brand": "Sample Brand",
  "marketplace": "com",
  "marketplaceDomain": "www.ebay.com",
  "price": 100.0,
  "priceMax": 150.0,
  "priceText": "$100.00 to $150.00",
  "currency": "USD",
  "originalPrice": 125.0,
  "discountPercent": 20.0,
  "listingType": "buy_it_now",
  "bidCount": null,
  "timeLeft": null,
  "shippingCost": null,
  "shippingText": "Free delivery",
  "freeShipping": true,
  "itemLocation": "United States",
  "returnsText": "Free returns",
  "freeReturns": true,
  "soldCount": 42,
  "sellerName": "examplestore",
  "sellerUrl": "https://www.ebay.com/sch/examplestore/m.html",
  "sellerFeedbackPercent": 98.5,
  "sellerFeedbackCount": 1234,
  "badges": ["Sample Badge"],
  "imageUrl": "https://i.ebayimg.com/images/g/000000000000/s-l500.webp",
  "attributes": ["Buy It Now", "Free delivery", "Located in United States", "Free returns", "42 sold"],
  "images": ["https://i.ebayimg.com/images/g/000000000000/s-l1600.jpg"],
  "itemSpecifics": { "Processor": "Sample Processor", "SSD Capacity": "1 TB", "Seller Notes": "Sample seller notes." },
  "categoryPath": ["Electronics", "Computers", "Laptops"],
  "description": "Full seller description text appears here when Read item pages is on.",
  "sellerNotes": "Sample seller notes.",
  "mpn": "SAMPLE-MPN",
  "model": "Sample Model",
  "color": "Black",
  "gtin": null,
  "listPrice": 125.0,
  "availability": "In Stock",
  "quantityAvailable": 5,
  "quantitySold": 42,
  "paymentMethods": ["PayPal", "Visa"],
  "shippingDetails": "Shipping: Sample shipping line",
  "returnPolicy": "Returns: 30 day returns accepted",
  "listedAt": "2026-01-01T00:00:00Z",
  "updatedAt": "2026-01-01T00:00:00Z",
  "searchQuery": "laptop",
  "sourceUrl": "https://www.ebay.com/sch/i.html?_nkw=laptop",
  "pageNumber": 1,
  "scrapedAt": "2026-01-01T00:00:00Z",
  "changeType": "NEW",
  "changedFields": [],
  "firstSeenAt": "2026-01-01T00:00:00Z",
  "lastSeenAt": "2026-01-01T00:00:00Z"
}

Plan Requirement

A paid Apify plan with residential proxies is required for reliable results. In testing, eBay answered datacenter connections with an HTTP 403 "Error Page | eBay" and no listings, while residential connections loaded pages (sometimes after one refused first attempt in the same session, which the actor retries). The Proxy & connection section therefore defaults to residential.

Free Apify plans do not include residential proxies. On those plans the run continues on a backup connection with a note in the log, but pages may still be unavailable. If no page loads, the run fails with: "No results: residential proxies are unavailable on this run (Apify free plans do not include them), and the backup connection could not load the requested eBay pages. Upgrade at https://apify.com/pricing or supply your own proxy URLs, then try again." If the proxy itself refuses the account, the message says your Apify plan does not include proxy access. Any other total failure reads "Could not load any results page (the requests failed), so nothing was collected. This is a connection problem, not an empty search." You can also supply your own proxy URLs. No eBay account, API key or other credential is required. The Pricing tab shows the current per-event rates for this actor.

FAQ

How much does it cost?

This actor uses Apify's pay-per-event pricing: one event per listing returned, one extra event per listing whose item page you had read, and one one-time event when the run starts. The Pricing tab shows the current rates. Use Max listings to cap the cost of any run.

Is it legal to scrape eBay?

This actor collects only publicly available listing data. You are responsible for how you use it: follow eBay's terms and the laws that apply to you, and get legal advice if you plan commercial redistribution. Listing facts such as price and availability are generally not protected, but photos and seller-written descriptions may be subject to third-party rights.

Can I get only new or changed listings on a schedule?

Yes. Schedule the actor from the Schedules tab and turn on Incremental mode. Each run then returns new and updated listings; with Emit expired on, a complete scan also returns expired listings, and a listing that later comes back is returned as reappeared. Unchanged listings are suppressed and not billed by default.

Why can't I filter for sold or completed listings?

Both eBay's LH_Sold=1 and LH_Complete=1 search filters were tested live and answered with eBay's sign-in page instead of results: those views sit behind an eBay account and cannot be read without one. This scraper does not offer a sold-listings toggle rather than offer one that would silently return nothing. If a pasted link carries either filter and hits the sign-in page, the log warns that the address was answered with a sign-in page; the run still finishes, with no rows from that link.

Why did my run fail instead of returning an empty dataset?

If every result page was refused, the run stops with a clear message so "no listings matched" is never confused with "nothing could be read". Run it again in a few minutes; on a free plan, supply your own proxy URLs or upgrade (see Plan Requirement). A search that genuinely matches nothing still succeeds with zero rows.

Can I use it with AI agents or MCP?

Yes. Call it from any Apify integration or MCP client, and use the connector field to push results into Notion, Linear or Airtable.

🔗 More scrapers you might like

Pair this actor with these related scrapers from the same team:

🛒 MUJI Scraper Scrape MUJI products from the United States, Canada and Australia storefronts. Search by...	🏷️ Vinted Multi-Country Scraper Scrape Vinted across 27 country marketplaces: keyword search with brand, price and...
🛒 Depop Scraper Scrape Depop search results, listings, shops, and reviews to JSON, CSV, or Excel. Extract...	🚗 Trade Me NZ Scraper Scrape trademe.co.nz across all five sections: Marketplace goods and auctions, Property...
⚽ Whatnot Scraper Scrape Whatnot live and upcoming shows, queued items and lots, and seller profiles...	🛒 UNIQLO Scraper Scrape UNIQLO products across 21 country storefronts in local currency. Price with pre...

👉 Browse all abotapi scrapers

💬 Support

• 🐞 Found a bug or a missing field? Open a ticket on the Issues tab. We usually reply within hours.
• 🛠️ Need another site, extra fields or a private build? Email contact@abotapi.com.
• ⭐ Enjoying it? A quick review on the actor page helps other users find it.