DEV Community

James
James

Posted on

Unlimited eBay Sold Data API (2026): One Flat Price, No Quota

Yes, there is an eBay sold data API with unlimited requests: the eBay Sold Listings API has an Unlimited plan with no monthly request cap, for one flat monthly price. Each request returns up to 240 completed eBay sales across 8 eBay marketplaces, and sold search, active listings, accepted Best Offer prices, batch jobs and CSV export are all included. It is the only eBay sold data API with an unlimited plan, and because the price is flat, every extra request you make lowers your cost per request.

Last checked: November 2026.

"Unlimited" is a word that gets stretched in API pricing, so this post sticks to what it means in practice: what you stop worrying about, how to size a workload yourself, and how to write a client that never needs to count requests. Current prices are on the pricing page.

What "unlimited" changes in practice

On a quota plan, every new user, SKU or keyword is a line item. You budget requests, alert at 80%, and hope a busy week does not stop your app before the period resets. The Unlimited plan removes that whole layer:

  • No quota to manage. There is no monthly allowance to run out of, so nothing stops mid-month.
  • No per-request accounting. Responses on Unlimited leave out the X-Usage-Remaining header, and GET /account/usage returns quota and remaining as null. There is nothing to count down.
  • More usage does not change the bill. Twice the users or twice the catalog costs the same.
  • Same key, same code. Moving to Unlimited is a change in the dashboard. Endpoints, fields and your API key stay the same.
curl "https://api.ebaysoldlistingsapi.com/account/usage" \
  -H "Authorization: Bearer YOUR_API_KEY"
# on Unlimited: "quota": null, "remaining": null
Enter fullscreen mode Exit fullscreen mode

Who actually needs unlimited eBay sold data

  • Apps with users. A price checker, collection tracker or reseller tool where every user search is an API call.
  • Card shops and resellers refreshing comps on a whole inventory every day.
  • Repricers watching thousands of SKUs across several eBay sites.
  • Data teams building price indexes or training sets from completed sales.

What these have in common: request volume is driven by things you do not control (users, inventory size), so a cap is a risk, not a budget.

Size your workload yourself

You do not need a pricing calculator to know where you stand. One request is one search: one keyword, one page, up to 240 sales. So:

def monthly_requests(keywords, pages_per_keyword=1, refreshes_per_day=1, days=30):
    return keywords * pages_per_keyword * refreshes_per_day * days


print(monthly_requests(500))                          # 500 items refreshed daily
print(monthly_requests(2_000, pages_per_keyword=2))   # deep comps on 2,000 items
print(monthly_requests(50, refreshes_per_day=24))     # 50 hot items, hourly
Enter fullscreen mode Exit fullscreen mode

Add your app's user searches on top (users times searches per user per month). Then compare the total with the plans on the pricing page. If the number is large, growing, or hard to predict, Unlimited is the plan that takes the question away. Only successful requests count; errors are never billed, so retries do not inflate the number.

Bulk workflows: where unlimited pays off

Unlimited is at its best with bulk tools, because you stop thinking about how many searches a job "costs".

Paging for deep comps. Every response has count, page, hasNextPage and totalResults. Keep requesting page=2, page=3 while hasNextPage is true:

curl "https://api.ebaysoldlistingsapi.com/scrape?keyword=pokemon+booster+box&count=240&page=2" \
  -H "Authorization: Bearer YOUR_API_KEY"
Enter fullscreen mode Exit fullscreen mode

Batch jobs for catalogs. POST /scrape/batch takes up to 100 keywords per job and runs them asynchronously. One batch runs per account at a time; polling and the CSV download are free.

curl -X POST "https://api.ebaysoldlistingsapi.com/scrape/batch" \
  -H "Authorization: Bearer YOUR_API_KEY" -H "Content-Type: application/json" \
  -d '{"keywords": ["ps5 slim", "steam deck oled", "nintendo switch oled"], "ebaySite": "ebay.com"}'

# when the job reports "complete":
curl "https://api.ebaysoldlistingsapi.com/scrape/batch/JOB_ID/download.csv" \
  -H "Authorization: Bearer YOUR_API_KEY" -o catalog.csv
Enter fullscreen mode Exit fullscreen mode

CSV on any search. Add format=csv and pipe results straight into a warehouse, Excel or Google Sheets.

Every marketplace. ebay.com, ebay.co.uk, ebay.de, ebay.com.au, ebay.ca, ebay.fr, ebay.it and ebay.es, each in local currency, through the ebaySite parameter.

A client that never counts requests

With no quota, the only thing a client has to respect is the rate. The cleanest way is an adaptive pacer: go fast, back off when the API says so (429 with Retry-After), and speed up again slowly after successes. No hard-coded plan numbers, so the same code works on any plan.

import os
import time
import requests

API = "https://api.ebaysoldlistingsapi.com/scrape"
HEADERS = {"Authorization": f"Bearer {os.environ['EBAY_SOLD_API_KEY']}"}


class AdaptivePacer:
    """Additive increase, multiplicative decrease: learns the safe pace by itself."""

    def __init__(self):
        self.gap = 0.0                      # seconds between calls, starts with none

    def before(self):
        if self.gap:
            time.sleep(self.gap)

    def ok(self):
        self.gap = max(0.0, self.gap - 0.02)

    def pushed_back(self, retry_after):
        self.gap = min(10.0, max(self.gap * 2, 0.25))
        time.sleep(retry_after)


pacer = AdaptivePacer()


def sold(keyword, **params):
    for attempt in range(6):
        pacer.before()
        r = requests.get(API, params={"keyword": keyword, **params}, headers=HEADERS, timeout=90)
        if r.status_code in (429, 502, 503, 504):
            pacer.pushed_back(int(r.headers.get("Retry-After", 2 ** attempt)))
            continue
        r.raise_for_status()
        pacer.ok()
        return r.json()
    r.raise_for_status()


for kw in ["steam deck oled", "ps5 slim", "shure sm7b"]:
    data = sold(kw, bestOfferPrice="true")
    print(kw, data["count"], "sales, more pages:", data["hasNextPage"])
Enter fullscreen mode Exit fullscreen mode

Notes:

  • Use a 90 second client timeout. Most searches return in a few seconds; a hard one can take longer.
  • Retry 429, 502, 503 and 504. Do not retry 400 or 401; they need a fix.
  • bestOfferPrice=true adds acceptedOfferPrice on sales that closed through an accepted Best Offer, at no extra cost.

What is included

Everything the API does, with no add-ons:

  • Sold and active searches (sold=false), up to 240 results per request.
  • The real accepted Best Offer price with bestOfferPrice=true.
  • Batch jobs of up to 100 keywords, with a CSV download.
  • All 8 eBay marketplaces, in local currency.
  • JSON or CSV on any search.
  • The hosted MCP connector for Claude, Cursor and other AI clients.

FAQ

Is the Unlimited plan really unlimited?
There is no monthly request cap. Like every API, requests have a per-minute rate limit, and the client above adapts to it on its own.

Why is a flat price the cheapest option at high volume?
Because your cost per request keeps falling as you use more. On quota plans it stays fixed and you buy the next tier when you grow.

Does Unlimited include Best Offer prices, batch jobs and CSV?
Yes, along with active listings and all 8 eBay marketplaces.

Can I try the API before choosing Unlimited?
Yes. You can start free with no card, test your real keywords, and upgrade from the dashboard later. Your key stays the same.

Do other eBay sold data APIs offer unlimited requests?
Not as of our comparison: the others sell monthly quotas, credits or per-result pricing.

Get started

Create a free key at ebaysoldlistingsapi.com, run your heaviest keywords through the client above, then see the Unlimited plan on the pricing page. Our unlimited API comparison shows how each provider bills, and the batch jobs docs cover bulk runs.

Top comments (0)