Amazon Keyword Research Tool

Find the search terms Amazon itself puts in front of shoppers. Give the actor a few seed keywords and it expands each one into the terms Amazon suggests as people type — the same drop-down every shopper sees — then optionally measures how contested each term is by looking at the products Amazon actually returns for it. Built for sellers writing listings, agencies planning campaigns, and anyone who would rather work from Amazon's own suggestions than from guesswork.

What you can do with it

• Build a keyword list for a listing. Expand your product term through the alphabet and take the long-tail phrases shoppers actually type.
• Find low-competition entry points. Sort by competition and look for terms where few paid placements and shallow result sets suggest room to rank.
• Research a niche before you enter it. Price ranges and the brands that keep appearing tell you what a category costs to compete in.
• Plan advertising. Every keyword comes back with how much of Amazon's first page of results is paid placement.
• Discover the words shoppers use for your product. Suggestions surface the phrasing and brand modifiers real buyers type, which is rarely the wording on the box.
• Repeat it per marketplace. The same seed produces a different list on each country's Amazon site.

What you get

One row per discovered keyword. Without competition data, a row looks like this:

{
  "keyword": "laptop stand",
  "seedKeyword": "laptop",
  "source": "suggestion",
  "depth": 1,
  "expandedFrom": "laptop",
  "suggestionRank": 3,
  "suggestionType": "KEYWORD",
  "marketplace": "us",
  "department": "aps",
  "hasCompetitionData": false,
  "scrapedAt": "2026-08-07T12:00:00.000Z"
}

With Include competition data switched on, each row also carries what Amazon serves for that term:

{
  "keyword": "laptop stand",
  "seedKeyword": "laptop",
  "suggestionRank": 3,
  "hasCompetitionData": true,
  "resultCount": 48,
  "resultPagesAvailable": 7,
  "sponsoredCount": 6,
  "organicCount": 42,
  "minPrice": 9.99,
  "maxPrice": 199.99,
  "medianPrice": 34.99,
  "currency": "USD",
  "averageRating": 4.55,
  "medianRatingsCount": 2320,
  "topBrands": ["HUANUO", "Nulaxy", "Lamicall"],
  "brandCount": 12,
  "competitionScore": 90,
  "competitionLevel": "high"
}

competitionLevel follows one published rule: how far Amazon's result set goes for the term, how much of the first page is paid placement, how established the products ranking for it already are, and how well they are rated. It is an indication, not a measurement of demand.

Input reference

Input	Type	Default	What it does
Seed keywords	list of text	wireless earbuds, yoga mat	Starting terms. Each is expanded into the suggestions Amazon offers shoppers typing it.
Expansion depth	choice	Alphabet expansion (a-z)	Suggestions only asks once per seed. Alphabet adds a–z to every seed. Shopping modifiers adds words like best, cheap, for, with, near and under, both before and after the seed. Deep does both, then expands the best results one level further.
Maximum keywords per seed	number	200	Stop after this many keywords discovered from each seed. Up to 3,000.
Department	choice	All departments	Limit suggestions to one Amazon department, as a shopper would with the search drop-down.
Include competition data	true / false	false	Run a real Amazon search for every keyword and attach the competition picture. Much slower, and billed separately.
Amazon marketplace	choice	United States	Which country's Amazon site to read. Prices come back in that marketplace's own currency.
Parallel targets	number	4	How many terms to work on at the same time, up to 10. Higher is faster, but Amazon serves fewer complete results when pushed hard.

Output fields

Field	Type	Description
keyword	text	The discovered search term.
seedKeyword	text	The seed it came from.
source	text	seed, suggestion, alphabet, modifier or deep — how the term was reached.
depth	number	0 for your own seed, 1 for its suggestions, 2 for the second level in deep mode.
expandedFrom	text	The exact term that was typed to surface it.
suggestionRank	number	The position Amazon gave the term in its own suggestion list. This is the only popularity signal Amazon publishes.
suggestionType	text	The kind of suggestion Amazon returned.
marketplace	text	The marketplace the run used.
department	text	The department the suggestions were scoped to.
hasCompetitionData	true / false	Whether the competition fields below were measured.
resultCount	number	How many products Amazon served for the term. Null when competition data was not collected.
resultPagesAvailable	number	How far Amazon's result set goes for the term.
sponsoredCount	number	How many of the served products were paid placements.
organicCount	number	How many were not.
minPrice / maxPrice / medianPrice	number	Price range across the served products.
currency	text	The currency those prices are actually in.
averageRating	number	Average star rating of the unpaid results.
medianRatingsCount	number	Middle review count of the unpaid results — how established the competition is.
topBrands	list	The brands appearing most often across the served products.
brandCount	number	How many distinct brands appeared.
competitionScore	number	0–100, from the rule described above.
competitionLevel	text	low, medium, high, or none when Amazon served nothing.
scrapedAt	text	Collection timestamp, in UTC.

Pricing

You pay per result, with no monthly minimum.

What you pay for	Price
Each keyword discovered	$0.40 per 1,000
Each keyword with competition data attached	$1.50 per 1,000 extra

Worked example. Two seeds with alphabet expansion typically yields around 250 keywords: $0.10. Switch on competition data for the same run and it is 250 × ($0.0004 + $0.0015) = $0.48.

Keyword discovery is deliberately cheap — it is the light half of the job. Competition data means a real Amazon search behind every single keyword, which is the heavy half, so it is optional, off by default, and priced on its own.

Limits & what this actor cannot do

• Amazon does not publish search volume for its suggestions, and this actor does not invent one. There is no monthly-searches number anywhere in what Amazon returns. What you get instead is suggestionRank: the order Amazon itself put the terms in, which is a genuine popularity signal and the only one that exists. Any tool quoting an exact Amazon search volume is modelling it, not reading it.
• Amazon's own printed result total is not reliable — the same keyword read minutes apart can print wildly different totals. Competition figures here are built only from products Amazon actually served, and competitionLevel is described as an indication rather than a count.
• Amazon returns about ten suggestions per term typed. That ceiling is exactly why expansion exists: alphabet and modifier modes ask many times to build a large list from a small seed.
• Amazon publishes a limited number of public results for any one search, so the competition picture reflects the results Amazon makes available rather than an unlimited catalogue.
• Brands are taken from the products Amazon serves for the term, so they reflect who is ranking today rather than everyone selling in the category.
• Suggestions and results are a snapshot at the moment of collection and keep changing afterwards.
• Suggestions, prices and competition differ between Amazon marketplaces; results reflect the marketplace you select.
• Sponsored placements appear inside Amazon's own results; they are counted separately so you can see how much of a term is paid.
• Speed depends on the size of the job and on Amazon's own response times. No fixed throughput is promised.
• Amazon's terms govern automated access. You are responsible for using the data lawfully, in line with the source site's terms, and for handling any personal data in line with applicable privacy law.

FAQ

Do I need an Amazon account? No. The actor reads only what Amazon shows the public.

Does it need my login or password? No. There is nothing to connect and nothing to sign in to — enter your seed terms and press start.

Can I schedule it? Yes. Suggestions shift with the season and with demand, so a weekly or monthly run gives you a useful history. Every row is stamped with its collection time.

Does it give me search volume? No, and neither does Amazon. Amazon publishes its suggestions in ranked order and nothing else, so this actor reports that rank honestly instead of printing a number that would be made up.

How many keywords will I get from one seed? Alphabet expansion typically produces somewhere between 100 and 250 keywords per seed; deep expansion produces more. Duplicates are removed across the whole run, so you are never charged twice for the same term.

Is the data complete? It is complete for what Amazon suggests: every term Amazon offered for the prefixes that were tried, de-duplicated. It is not every keyword in existence, because Amazon does not publish such a list.

Why do two of my keywords have the same competition level? Because the level is a band, not a score out of a thousand. Use competitionScore, sponsoredCount and medianRatingsCount when you need to separate terms inside the same band.