Try any of 1536 endpoints — live.

Look up MANY eBay listings in ONE call by item number — up to 15 per request → the same full record `item_detail` returns for each one (title, price, condition, brand/model/MPN, images, category path, item specifics, seller reputation, shipping terms, return policy, auction fields, quantities). Built for enriching a price list or a competitor's catalogue in one round trip instead of one item at a time. COST, stated plainly so you can budget it: eBay publishes no batched item endpoint — every listing is its own page — so this costs ONE UPSTREAM REQUEST PER ITEM ID, exactly as if you had called item_detail N times, and it is priced that way. What you save is wall-clock, not requests: the fetches run concurrently, measured at 25.6s sequential vs 4.1s for 8 ids (6.2x). `meta.upstream_calls` always reports the true number. Anything eBay does not resolve comes back in `not_found` rather than being silently dropped, so a batch of 15 can never hide which one went missing. If you only need card-level fields (item_id, title, price, condition, shipping, image, url) there is a genuinely cheaper route that costs ONE upstream call: `search` with query `(id1,id2,id3)` — eBay's own OR syntax — measured to resolve 6 ids per call (a 7th is dropped) and to return no item specifics, seller or return policy.