Real eBay sold prices as JSON: up to 240 sales per GET request, up to 1,095 days of price history, and the price a Best Offer actually closed at. One key, no approval process.
Not a developer? Use it in Google Sheets instead.
The same two calls now return sold prices from Reverb, StockX, GOAT, Grailed, Poshmark, Mercari, Vinted, Goldin, Fanatics Collect, MySlabs, Pristine and TCGplayer. Add source=stockx to read one marketplace, or source=all to merge all 12 in one call, with every sale tagged by where it sold. /v1/active returns what is for sale right now on the same 12. PSA's auction price record is its own source.
Add exclude=ai_junk and our AI model takes lots, parts, accessories and replicas out before the median. Every plan includes all of it, Free too, and one call is one request.
Plain HTTPS and JSON. No SDK required, no OAuth dance, no eBay developer approval.
A request is one search page of up to 240 sales, from eBay US with matching sales from other eBay sites merged in. Every plan includes real Best Offer prices, every filter and graded card parsing with grade and grader per sale. Failed requests are never counted.
A real key to try every filter on. No card, no expiry.
For one tool or a side project that prices daily.
For a product that prices continuously and pages deep.
For teams pricing whole categories at volume.
For a catalogue wide repricer running every day.
For a marketplace or data product built on comps.
For high volume pricing at platform scale.
Every plan pauses once its monthly quota is used and starts again when the UTC month resets. There is no overage charge. Upgrading raises the quota straight away and is prorated, and you can downgrade or cancel anytime.
Best Offer hides the real price. When a buyer and seller agree on an offer, the public sold page keeps showing the asking price. On high ticket items the gap often runs 20 to 40 percent, and anything priced from it inherits the error.
One parameter fixes it. Pass best_offer=1 and every sale carries a flag, with price set to the amount the sale actually closed at. Pass best_offer=only to get accepted offers and nothing else.
The public sold search stops at 90 days. CompsAPI goes back 1,095, so you price by trend instead of by snapshot.
Seasonality needs a full year. Sports cards climb into the playoffs, consoles spike in December, golf clubs peak in spring. A 90 day window hides every one of those curves and quietly misprices anything seasonal.
Depreciation needs three. The chart above is real API output: iPhone 14 Pro average sold price, down 30 percent across 24 months. Price trade ins, buybacks and insurance values from the curve, not the moment. Pass days=1095 and page through it.
GET /v1/soldGET /v1/compsGET /v1/activeGET /v1/usagetitleitem_idurlpricebest_offerlast_soldformatshipping_costp50 1.0 s · p95 about 2 s · failed requests never counted · time based paging, no duplicates between pages
Check what an item actually sells for before you buy it. Pull eBay comps by condition, see the spread between asking and accepted prices, and price listings to move instead of sit.
Reseller pricing ›Graded card comps for PSA, BGS and CGC with grade and language filters. Three years of history shows how a card priced across whole seasons, not one hot week.
Trading card comps ›High ticket sales close as Best Offers, so public sold prices run high. Price watches, jewelry and designer pieces from accepted offers, not wishful asking prices.
Watch and luxury comps ›Feed live sold comps into your repricer, valuation model or buy box logic. Time based paging keeps syncs clean and /v1/usage makes metering your own customers simple.
Instant valuations for trade in flows, consignment intake and buyback offers, backed by the prices buyers actually paid on eBay rather than list prices.
The price history API ›Monthly average prices and sales volumes per query across 36 months. Track launch cliffs, seasonality and depreciation curves product by product.
Market research ›Same data, no API calls. Type an item into a Google Sheet and get the real median sold price back in the cell. Free template, copy it and add your key.
=COMPS_MEDIAN("iphone 14 pro 256gb") returns the median of real completed sales. Lots, bundles and wrong models are filtered out before the median is worked out.
Median, how many sold, the typical price band and a table of the newest sales with dates. Filter by condition, or look back up to three years with one extra argument.
The sheet uses the price a Best Offer actually closed at, the number eBay's own sold page hides behind a crossed out asking price. Price a lot of items in one pass.
Free template · works with the same free key · 100 searches a month, no card
findCompletedItems is retired. eBay switched off the Finding API call developers used for sold and completed listings. CompsAPI returns the same workflow's data, keyword search over sold items with price, date, format and shipping, as one GET request.
Read the migration guide ›Marketplace Insights is closed. eBay's Marketplace Insights API is limited to approved partners and closed to new applicants. CompsAPI needs no approval, no eBay developer account and no quota negotiation. Start with a free key today.
Read the comparison ›Scrapers hit the 90 day wall. Scraping sold search means login walls, 90 days of history, asking prices on Best Offer sales and constant maintenance. An API that maintains itself costs less than a scraper that doesn't.
No. CompsAPI is an independent service and is not affiliated with eBay. eBay's own sold data options are the Marketplace Insights API, which is closed to new applicants, and the retired findCompletedItems call (every route, in detail). CompsAPI is a drop in replacement for both workflows: one GET request, an API key, no approval process.
Pass days up to 1095 on /v1/sold. The response pages through up to three years of sold history for any search, with the sold date and time on every row.
Yes. With best_offer=1, every sale carries a best_offer flag and price is the amount the sale actually closed at, in every category. In our October 2026 verification, all 49 accepted-offer sales in the set carried the asking price on the public sold page and the real price here.
CompsAPI covers the findCompletedItems workflow: keyword search over sold listings with price, sold date, format and shipping per item, returned as JSON with up to 240 rows per page.
Yes. Marketplace Insights is limited to approved eBay partners. CompsAPI returns sold listings data with no approval process, a free tier to test on and up to 1,095 days of history instead of 90.
No. The free Google Sheets template puts the same data in a spreadsheet cell: type an item, get the real median sold price back. It uses the same free key and the same 100 searches a month. The API is there when you want to automate it.
A comp, short for comparable sale, is a recent real sale of the same item in similar condition. Resellers price inventory from comps instead of list prices. /v1/comps automates it: a cleaned median, the spread and the newest matching sales in one call.
Sales usually appear within minutes. Each response carries the sold date and time, not just the date.
eBay US is primary, with matching sales from other eBay sites merged in. The API can also pin any one of 8 eBay sites: US, UK, DE, FR, IT, ES, CA and AU.
One search page of up to 240 sales. /v1/usage is free to call and shows your quota in real time. Failed requests are never counted.
Paid plans run at up to 120 requests a minute. For more throughput, bulk history exports or an SLA, ask about Enterprise.
Both. Every row carries format, AUCTION or FIXED_PRICE, and the condition and Best Offer filters work across both formats.
Yes. The Free tier is 100 requests a month with every filter enabled, no card. Create a free account and your key is on screen in under a minute.
100 free requests a month · every filter on · no card
Get a free key