Quickstart
Get a Soldgraph API key, send your first eBay sold listings search, read the median price, and price a whole list in one call.
#Get a key
Create an account and create a key in API keys. You start with 100 free requests, no credit card required. Save the key as SOLDGRAPH_KEY in your environment. The key is shown once.
Is an AI assistant writing your code? Give it the ChatGPT prompt and skip to running the result.
#Search eBay sold listings
curl -G https://api.soldgraph.com/v1/ebay/sold \ -H "Authorization: Bearer $SOLDGRAPH_KEY" \ --data-urlencode "q=sony wh-1000xm5" \ -d page=1 \ -d count=200#Get the result
The first call for a search usually returns 202 with status=pending and a poll_url. Poll it with the same key and ?wait=20 until status is complete or failed:
curl "https://api.soldgraph.com/v1/jobs/REQUEST_ID?wait=20" \ -H "Authorization: Bearer $SOLDGRAPH_KEY"poll_url is a path, so join it to https://api.soldgraph.com. If someone ran the same search in the last 15 minutes, you get 200 and the result right away. That still costs one request, and so does a search that finds nothing. Only failed searches are free.
Search jobs explains the full flow, retries and pagination. For Python and Node clients that handle polling and retries for you, see the examples on GitHub.
#Example response
A complete response looks like this. The data is an example.
{ "request_id": "00000000-0000-4000-8000-000000000001", "status": "complete", "credits": 1, "cached": false, "result": { "provider": "ebay", "country": "us", "query": "sony wh-1000xm5", "page": 1, "page_size": 200, "count": 1, "reported_total": 1, "next_page": null, "next_cursor": null, "collected_at": "2026-09-23T14:02:11Z", "schema_version": 2, "completeness": "provider_page_only", "summary": { "priced_count": 1, "currency": "USD", "median": 189, "low": 189, "high": 189, "p25": 189, "p75": 189, "earliest_sold_date": "2026-09-21", "latest_sold_date": "2026-09-21", "best_offer_accepted_count": 0, "median_excluding_best_offer": 189 }, "data": [ { "id": "123456789012", "title": "Example eBay listing", "link": "https://www.ebay.com/itm/123456789012", "format": "fixed_price", "displayed_price": { "amount": 189, "currency": "USD" }, "displayed_price_text": "$189.00", "displayed_price_range": null, "best_offer_accepted": false, "displayed_shipping": { "amount": 9.99, "currency": "USD" }, "shipping_text": "+$9.99 delivery", "sold_date": "2026-09-21", "sold_date_text": "Sep-21 14:35", "condition": "Pre-Owned", "bid_count": null, "seller_text": "sonyfan_outlet (2,418) 99.6%", "location_text": "Located in United States", "image": "https://i.ebayimg.com/images/g/example/s-l500.jpg" } ], "field_notes": "Fields that could not be read are null. Displayed prices are not verified accepted-offer prices." }}#Read the price
Every listing search includes a summary, so you get a number without writing any maths:
{ "priced_count": 200, "currency": "USD", "median": 189, "low": 120, "high": 260, "earliest_sold_date": "2026-09-02", "latest_sold_date": "2026-09-21", "best_offer_accepted_count": 9 }It is plain maths over the rows on the page, with nothing removed. The search above asked for count=200, the most eBay gives in one request; leave count out and you get 40.
#Price a list
To price many items, send them as one batch. You get one ID to poll and a summary for each item.
curl -X POST "https://api.soldgraph.com/v1/batches" \ -H "Authorization: Bearer $SOLDGRAPH_KEY" \ -H "Content-Type: application/json" \ -d '{ "defaults": { "endpoint": "ebay/sold", "count": 120 }, "searches": [ { "q": "sony wh-1000xm5" }, { "q": "bose qc45" }, { "q": "airpods max" } ] }'Then poll /v1/batches/BATCH_ID?wait=20. Add ?format=csv to download every row as one file.
Overviews: eBay sold listings API · findCompletedItems alternative · Poshmark sold listings API · Mercari sold listings API · Pokémon card price API · Sold comps API for reseller apps