Amazon Sales Estimate API for Any ASIN or BSR

What this API returns

Amazon does not publish how many units a product sells. Every sales figure you see in a seller tool is an estimate, modelled mainly from Best Sellers Rank. This API gives you those estimates in three shapes:

A real request for one cat water fountain in the US (ASIN B08NCDBT7Q) returned a daily series from 1 September 2025 to 14 October 2026 and fourteen monthly rows. Part of it:

Daily rows carry bsr next to sales, so you can see the input beside the estimate. The series shows what a monthly average hides: Black Friday at more than three times a normal day, a summer month half again above September, Christmas Eve at a third of a normal day.

Request example

curl --request POST \
  --url https://ecommercedataapi.com/v1/amazon/sales/prediction/asin \
  --header "Content-Type: application/json" \
  --header "X-API-Key: ${ECOMMERCE_DATA_API_KEY}" \
  --data '{
    "marketplace": "US",
    "asin": "B08NCDBT7Q"
  }'

• asin is required; marketplace defaults to US. Supported codes: US, UK, DE, FR, IT, ES, MX, CA, AU, JP, IN, AE, BR.
• The response holds asinDetail (title, category, rating), dailyItemList and monthItemList. The current month is partial — October 2026 above held 14 days.
• One call returns the full history, so there is no paging to pay for.

From a rank to a sales figure

When you only have a rank — from a screenshot, a category list or your own tracking — the BSR endpoint converts it. In Pet Supplies (categoryId 2619533011), BSR 300 returned estMonthSales: 23523 and estDailySales: 784, along with an itemList covering ranks 1 to 1,300 in the same call:

That is the curve every estimate is built on, and it is steep: moving from rank 1,000 to rank 100 is a fivefold change. The fountain above held BSR 240–430 through September and was estimated at 22,730 units — consistent with the curve.

Two estimates side by side

The ASIN sales trend endpoint uses a separate monthly model. For the same product, September 2026 came back as 22,548 parent units against 22,730 from the daily model — within 1%. July 2026 did not agree: 12,190 against 35,501. Two models agreeing is evidence; two models disagreeing is a reason to look at the daily series before you trust either month.

The trend endpoint also returns childUnitSales in round steps (10000, 9000, 8000). Read it as a band, not a count.

Common workflows

1. Size a competitor. Pull the monthly series for the top listings in a niche and add them up. Pair it with product research to find those listings first.
2. Rank a shortlist cheaply. If your shortlist already has BSR and category, one BSR call returns the whole rank curve, so you can estimate many products from a single request.
3. Spot seasonality. Daily rows show event spikes and holiday dips; monthly rows show the season. Seasonal product research covers how to read them.
4. Check a price change. price sits in the same row as sales, so you can see whether a price move came with a sales move. For longer price history, see price history.

Limits worth knowing

• These are estimates, not Amazon figures. Use them to compare products and see direction. What sales estimates can and cannot tell you explains the model and its blind spots.
• BSR curves are per category. Rank 300 in one category says nothing about rank 300 in another.
• Ranks are relative to the rest of the category, so a whole category slowing down can leave rank unchanged while sales fall.
• Each successful call consumes one billing unit. Handle 429 responses as described in errors and rate limits.

FAQ

Does Amazon offer an official sales API for other sellers' products? No. The Selling Partner API reports sales for your own account only. Sales for any other ASIN can only be estimated.

Why do tools disagree on the same ASIN? Each tool fits its own rank-to-sales curve. Compare tools on the same products over the same month, and prefer the one whose numbers move with events you can verify.

Can an AI agent call it? Yes. All three endpoints are tools on the remote MCP server, so an assistant can answer "how many does this ASIN sell" with data rather than a guess.