Google Trends API for Amazon Sellers

What this API returns

The Google Trends endpoint takes a keyword and an Amazon marketplace and returns how interest in that keyword has moved over time on Google. It is meant for one job: checking what happens outside Amazon before you decide something on Amazon data alone.

The response has four parts:

The documented link covers a five-year window (2020-09-14 to 2025-09-14) for geo=US, so one call returns several years of history. Both documented timestamps fall on a Sunday, which suggests weekly points when monthly is left at its default. The full field list is in the Google Trends reference.

Request example

curl --request POST \
  --url https://ecommercedataapi.com/v1/amazon/google/trends \
  --header "Content-Type: application/json" \
  --header "X-API-Key: ${ECOMMERCE_DATA_API_KEY}" \
  --data '{
    "marketplace": "US",
    "keyword": "iphone stand",
    "googleProp": "web",
    "monthly": false
  }'

• marketplace is required and accepts US, JP, UK, DE, FR, IT, ES, CA, IN, MX, BR, AU and AE.
• googleProp chooses the Google surface: web for web search, shoppingCart for Google Shopping search. Shopping search is usually closer to buying intent, so run both when a term matters.
• monthly defaults to false. Set it to true when you want monthly aggregation.
• The request body is validated strictly: unknown fields are rejected rather than ignored.

Common workflows

1. Confirm seasonality. A monthly history of Amazon searches from Keyword Research Trends says when a term climbs on Amazon. If the Google curve climbs in the same months, the pattern is sturdier. Seasonal product research shows how to do this before you pick a month.
2. Check a term is not a one-year spike. Several years on one curve separate a season that repeats from a fad that happened once.
3. Compare web and shopping interest. A term rising in Google Shopping before it rises on web search is worth a closer look.
4. Set a launch window. Decide the month you need stock in place from the month the curve starts to climb, minus your own restocking cycle.

Limits worth knowing

• The score is relative. Google scales each result so that its highest point is 100, so value is not a search count and not comparable between two separate requests. To compare two keywords, compare the shape of their curves, not their levels.
• One keyword per request. Call the endpoint once per term.
• Google interest and Amazon demand are related but not the same. Use this as a cross-check, not as a forecast of Amazon sales.
• Each successful call consumes one billing unit. Handle 429 responses as described in errors and rate limits.

FAQ

Is this Amazon data? No. It is Google search interest, returned through the same API and account as the Amazon endpoints so you can put the two side by side.

Can I compare several keywords in one call? No. Each request covers one keyword and is scaled on its own, so call once per term and compare the curves.

Can an AI agent call it? Yes. The endpoint is exposed as a tool through the remote MCP server, so an agent in Claude Code or Cursor can pull a curve and read it in the same conversation.