Keyword Autocomplete Scraper: Google, YouTube, Bing, Amazon avatar

Keyword Autocomplete Scraper: Google, YouTube, Bing, Amazon

Pricing

from $0.10 / 1,000 suggestions

Go to Apify Store
Keyword Autocomplete Scraper: Google, YouTube, Bing, Amazon

Keyword Autocomplete Scraper: Google, YouTube, Bing, Amazon

Autocomplete keyword ideas from Google, YouTube, Bing, DuckDuckGo and Amazon in one run: each keyword expanded a-z, with question and comparison modifiers, for 36 countries. Deduplicated, with position and relevance. $0.001 per keyword.

Pricing

from $0.10 / 1,000 suggestions

Rating

0.0

(0)

Developer

DTK Foundry

DTK Foundry

Maintained by Community

Actor stats

0

Bookmarked

2

Total users

1

Monthly active users

a day ago

Last modified

Share

Keyword ideas from the autocomplete boxes of Google, YouTube, Bing, DuckDuckGo and Amazon, in one run. Each keyword is widened the way keyword tools do it (keyword + a ... z, question words, "for / with / vs ..."), every suggestion is de-duplicated, and you get where it came from: what was typed, its position in the list and Google's relevance score. 36 countries, any language.

  • Many ideas per keyword. "standing desk" with all three expansions gave 619 unique ideas on Google, 475 on Bing, 408 on YouTube, 294 on DuckDuckGo and 282 on Amazon (live run, 2026-10-22), in under 5 seconds.
  • Reliable. Tested on 200 varied keyword lookups across all five engines and 10 countries: 5,400 of 5,400 requests answered, no retries needed (the only empty results were English keywords on Amazon UK/DE/FR/ES, where Amazon itself suggests nothing). Then 40 keywords x 5 engines with every expansion, as one run: 200 of 200 answered, 79,303 unique suggestions, 0 failed.
  • Five engines, one format. Pick any mix. Amazon uses the store of the country you choose (12 stores).
  • Clean keyword output. Google's website shortcuts ("http://...") are left out; every row is a search phrase people actually type.
  • $0.10 per 1,000 suggestions. Pay only for unique suggestions delivered, and cap how many you want per keyword. Keywords that fail, or return nothing, are never charged.

Input

FieldTypeDefaultDescription
keywordsarray[]One keyword or phrase per line. Empty = a free format example.
enginesarray["google"]Any of google, youtube, bing, duckduckgo, amazon.
expansionsarray["alphabet"]alphabet (keyword + a ... z), questions (how / what / why ... + keyword), prepositions (keyword + for / with / without / vs / near / to / like / or / and). Empty = the keyword alone.
maxSuggestionsPerKeywordinteger0Cap per keyword and engine (0 = no limit). The keyword's own suggestions come first, then each expansion in order.
countrystringusTwo-letter country code: us, gb, ca, au, nz, ie, in, sg, za, de, at, ch, fr, be, nl, es, mx, ar, co, cl, it, pt, br, pl, se, dk, no, fi, jp, kr, tr, id, ph, my, ae, sa.
languagestringcountry's ownTwo-letter language code for Google, YouTube and Bing (en, de, es ...).
outputFormatstringsuggestionssuggestions = one row per unique suggestion; grouped = one row per keyword and engine with a list.
{
"keywords": ["coffee maker", "standing desk"],
"engines": ["google", "youtube", "amazon"],
"expansions": ["alphabet", "questions"],
"country": "us"
}

Output (one record per unique suggestion)

{
"status": "ok",
"suggestion": "standing desk and walking pad combo",
"keyword": "standing desk",
"engine": "bing",
"query": "standing desk and",
"expansion": "prepositions",
"position": 10,
"relevance": 1291,
"timesSeen": 1,
"country": "us",
"language": "en",
"scrapedAt": "2026-10-22T17:15:57+00:00"
}
  • query is what was typed into the box and position the suggestion's place in that list (1 = top). expansion says which expansion produced it (keyword for the keyword itself).
  • relevance is the engine's own score where it gives one (Google, YouTube and Bing); higher = shown higher. Amazon and DuckDuckGo give null.
  • timesSeen counts how many of the keyword's lookups returned the same suggestion. Ideas seen in many lookups are the strongest.
  • Suggestions are de-duplicated per keyword and engine, ignoring case. The same idea from two engines appears once for each engine.
  • In grouped mode each record has suggestions (a list in first-seen order), count, queries and queriesAnswered.
  • A keyword with no suggestions gives one record with status: "no-results"; one that could not be looked up gives status: "failed" and an error. Neither is charged.

Pricing

Pay per event, no subscription.

EventPriceWhen it is charged
Suggestion$0.0001Once per unique suggestion delivered ($0.10 per 1,000), in either output format.

What a keyword typically costs on one engine (live run, 2026-10-22, US):

ExpansionSuggestions per keywordCost per keyword
Keyword alone10-15about $0.0013
Alphabet (a-z)Google ~250, Bing ~200, YouTube ~170, DuckDuckGo ~130, Amazon ~100$0.01-0.025
Alphabet + questions + prepositionsGoogle ~600, Bing ~490, YouTube ~425, DuckDuckGo ~280, Amazon ~185$0.02-0.06

Use maxSuggestionsPerKeyword to set a hard cost per keyword (100 = at most $0.01). Keywords with no suggestions or that could not be looked up are not charged. If a few expansion lookups fail after retries, you get (and pay for) only the suggestions that came back. Set a maximum charge in the run options and the actor stops before it would go over. Apify's standard run-start event adds $0.00005 per run.

Limits, stated plainly

  • These are autocomplete suggestions, not search volumes. They show what people type, in the order each engine ranks them, but not how often.
  • Question and preposition words are English. For other languages use the alphabet expansion, which works in any language.
  • Google may mix in ideas tied to the location of the server it is asked from (a city name). The country setting steers the answers but does not pin them to one city.
  • Amazon has no suggest service for India on this endpoint, so Amazon covers 12 stores: us, gb, de, fr, ca, es, it, nl, jp, mx, au, br.
  • Up to 1,000 keywords per run.
  • Public data only: no login, no personal data.

Privacy

Only the keywords you give are sent to the engines you choose. Results are stored only in your run's dataset. The actor runs with limited permissions and reads nothing else in your account.