How to Get eBay Sold Prices for Sports Cards

A search for a player can mix inexpensive base cards with rare parallels, autographs, graded singles, and lots. Even a precise-looking query can omit the one detail that makes the price meaningful. Write down the card's identity before deciding which sales belong in your comparison.

A query such as 2018 Panini Prizm Luka Doncic #280 PSA 10 is a useful starting point, but it still does not specify the parallel. Add the intended variation, then inspect the titles and photographs. Missing information is not proof of a base card.

Keep the grading company alongside the number. A listing condition such as Pre-Owned is not a PSA grade, and conditionId is not a card-grading scale. PSA's official grading standards describe its grades; the seller's generic condition label does not replace that evidence.

On eBay, search the full card description and select Sold Items. Completed listings alone can include unsold items, so make sure you are looking at sales. For a repeatable table and export workflow, open CompSniper's dashboard search:

1. Enter the card details and choose the marketplace you want to price against. A UK seller can choose ebay.co.uk; a US seller can choose ebay.com.
2. Leave AI Relevance enabled as an initial cleanup step. It is not a guarantee that every retained card has the exact parallel, grade, or serial run you need.
3. Run the search and inspect the returned rows, sale dates, condition, price, and shipping. Open questionable listings before treating them as exact comps.
4. Use Export CSV for a spreadsheet, or switch to JSON to inspect the response. The displayed summary describes the returned sample, not every sale on eBay.

If you are signed out, sign in or create an account first. The keyword is preserved, and opening the page does not run a paid search automatically. Review the query and press Search when ready.

In our August 31, 2026 fixed 100-product study, the query 2018 Panini Prizm Luka Doncic #280 PSA 10 returned 240 priced rows. All 240 remained after that study's relevance pass. The median was $238.75, but the middle half ranged from $202.50 to $900.00, and the highest displayed price was $8,600.

That wide spread is a reason to inspect the underlying mix, not proof that the card was worth $8,600 or that all retained rows were exact equivalents. The query did not pin down a parallel. The aggregate CSV cannot establish the reason for each outlier, and 61 retained rows carried a Best Offer accepted marker.

For a card-pricing app or repeatable inventory workflow, use CompSniper's eBay sold listings API. These examples request up to 20 rows for the same starting query. Set COMPSNIPER_API_KEY in your terminal or server environment first. Python requires the requests package; the JavaScript example runs on Node.js, not in a public browser.

curl --fail-with-body --max-time 75 --get \
  "https://api.compsniper.com/v1/scrape" \
  -H "Authorization: Bearer $COMPSNIPER_API_KEY" \
  --data-urlencode "keyword=2018 Panini Prizm Luka Doncic #280 PSA 10" \
  --data-urlencode "ebaySite=ebay.com" \
  --data-urlencode "sold=true" \
  --data-urlencode "count=20" \
  --data-urlencode "relevance=true"

The response contains listing rows in items and a price summary when available. A successful page uses one request regardless of whether it returns 5 or 20 rows. You can request up to 240 rows with count=240, but the number returned depends on available and retained matches. Choosing a larger count does not create more sales.

Always set ebaySite for your intended market. The REST API defaults to ebay.com; it does not infer your seller marketplace. Read the parameter reference for filters and the error guide before adding retries. For many cards, the Card Batch API provides a dedicated workflow.

Separate the item price from shipping and use the same basis for every comp. Do not combine currencies or compare a single card with a multi-card lot as if they were equivalent. A sold listing is marketplace evidence, not independent proof that payment settled or that the transaction was never reversed.

A Best Offer marker also matters. When eBay does not expose the accepted amount in the source listing, the visible amount may not be what the buyer paid. CompSniper's bestOfferAccepted flag highlights that limitation; null means the marker was not observed, not that the price has been independently verified.

eBay's Product Research is a separate first-party option with longer sales history and accepted Best Offer prices. Its card Price Guide also offers card-specific research. Those tools are not the same data-access contract as this REST endpoint.

Rare parallels and high-grade vintage cards may have very few recent sales. Start by checking the year, spelling, set name, card number, and marketplace. Remove a nonessential adjective before dropping the defining grade or parallel. If you compare a neighboring grade or variation, keep it in a separate group and label it as a related card, not an exact comp.

CompSniper's standard sold search covers recent source results, approximately the public 90-day window, not a guaranteed two-year archive. A small or empty result set can be the correct answer. Do not widen the query until unrelated cards produce a reassuring-looking average. Our sold-comp sample-size guide explains how to judge a small evidence set.

Written by CompSniper's founder. Source guidance checked September 12, 2026. The study uses CompSniper's own fixed research queries, not customer searches.