Google Autocomplete Scraper: keyword ideas from Google, YouTube, Amazon and Bing

Type a seed keyword, get the search suggestions people really see, from Google, YouTube, Amazon and Bing in one run. Expand a seed with a-z, question words or your own modifiers to turn one keyword into hundreds of long-tail ideas. Each row has the suggestion, the query that produced it, the source, country, language and rank. No login, no API key, no browser.

Price: $0.20 per 1,000 suggestions ($0.0002 each), no start fee. See Pricing below.

Why use this one

• Four engines, one schema. Google, YouTube, Amazon (12 marketplaces) and Bing, so you can compare what shoppers, viewers and searchers type.
• Expansion built in. alphabet (a-z, 0-9), questions (how, what, why, can...), modifiers of your choice, and recursionDepth 2 to feed suggestions back in.
• Country and language control, so results are repeatable for a market.
• You set the cost. maxSuggestions caps the run, duplicates are removed, and a rate limit is waited out instead of losing the keyword.
• Honest limit: autocomplete shows what engines suggest, not search volume, and each query returns about 10 suggestions, so volume comes from expansion.

Minimal input

{"keywords": ["standing desk"], "sources": ["google", "youtube"], "expansionMode": "alphabet"}

Sample output row (real platform run)

{
  "suggestion": "standing desk with drawers",
  "seedKeyword": "standing desk",
  "expansionQuery": "standing desk",
  "source": "google",
  "country": "US",
  "language": "en",
  "position": 3,
  "scrapedAt": "2026-10-11T07:07:47+00:00"
}

What does Google Autocomplete Scraper do?

For each seed keyword and each source it types the seed (and, if you choose, many variations of it) into the suggest box and collects what comes back, removing duplicates within each seed and source. Typical uses:

• SEO keyword research. Build long-tail keyword lists from one seed with alphabet and question expansion, without paying for a keyword tool.
• Content planning. The "how / what / why / can / is" questions people really type for a topic, ready to become blog posts, FAQs and video titles.
• YouTube and video SEO. What viewers search for on YouTube, which often differs from Google.
• Amazon product research. The shopping searches buyers type on Amazon in 12 marketplaces, for listings, ads and niche validation.
• Reputation and market monitoring. What Google and Bing suggest after a brand or competitor name, tracked on a schedule.
• AI agents and RAG. A structured "what do people ask about X" tool for agents and pipelines.

Pair it with Google Trends Scraper to see which of the suggested keywords are rising.

Input

Field	Type	What it does
keywords	list of text	Seed keywords, any language.
sources	list	Any of google, youtube, amazon, bing. Each source is queried for every keyword. Empty means Google only.
country	text	Two-letter country code, default us. Sets Google/YouTube gl, the Bing market and the Amazon marketplace (us, gb, de, fr, it, es, ca, jp, in, au, mx, br; other codes use amazon.com).
language	text	Language code, default en (de, ja, zh-TW ...). Sets Google/YouTube hl and the Bing market. Amazon follows the country.
expansionMode	choice	none (the seed only, about 10 suggestions), alphabet (seed + a-z, 0-9: 37 queries), questions (how, what, why, can, is, are, when, where, which, who, does, should, best before the seed; vs, versus, or, and, for, with, without, near me after it), modifiers (words before the seed).
modifiers	list of text	Words for modifiers mode. Empty means best, cheap, free, top, buy, new, online, diy.
recursionDepth	1 or 2	2 also types every suggestion found back in as a new query for many more long-tail results.
maxSuggestionsPerKeyword	number	Cap per seed and source, default 200. Also your cost cap.
maxSuggestions	number	Cap for the whole run, default 1000. The run stops there.
requestDelaySeconds	number	Pause between requests, default 0.3 s (randomised up to 1.4x).
proxyConfiguration	object	Optional Apify Proxy. Default is no proxy.

{
  "keywords": ["standing desk", "ergonomic chair"],
  "sources": ["google", "youtube", "amazon", "bing"],
  "country": "us",
  "language": "en",
  "expansionMode": "questions",
  "maxSuggestionsPerKeyword": 150,
  "maxSuggestions": 2000
}

Output

One dataset row per distinct suggestion.

Field	Meaning
suggestion	The suggested search phrase.
seedKeyword	The keyword you gave.
expansionQuery	What was actually typed to get it (the seed, standing desk a, how standing desk ...).
source	google, youtube, amazon or bing.
country, language	The codes used for the request.
position	Rank in the suggestion list for that query, 1 is the top.
scrapedAt	ISO 8601 UTC time.

{
  "suggestion": "standing desk converter",
  "seedKeyword": "standing desk",
  "expansionQuery": "standing desk",
  "source": "amazon",
  "country": "US",
  "language": "en",
  "position": 2,
  "scrapedAt": "2026-10-09T10:10:25+00:00"
}

The phrase you typed is never returned as its own suggestion. A suggestion found by several expansions is returned once, with the first query that produced it.

Pricing

$0.0002 per suggestion returned ($0.20 per 1,000), no start fee, billed through the pay-per-event event suggestion. Prices are final. Nothing is charged for queries that return nothing.

Worked examples:

• 1 keyword, Google only, no expansion: about 10 suggestions = $0.002.
• 1 keyword, Google, alphabet expansion: about 200 suggestions = $0.04.
• 20 keywords, all 4 sources, questions expansion, 100 per keyword and source: up to 8,000 suggestions = $1.60.

Set maxSuggestions and the run cannot cost more than that many times $0.0002. If you also set a maximum charge on the run, it stops there too.

Limits

• The suggest endpoints return about 10 suggestions per query (Bing up to about 25), so volume comes from expansion and recursionDepth.
• Suggestions depend on country and language, and Google may also personalise by location. Use country and language for repeatable results.
• Suggestions are what each engine shows today; they carry no search volume. Combine with a volume tool if you need numbers.
• Amazon suggestions come from the marketplace of the chosen country; unsupported countries fall back to amazon.com.

Use it from n8n, Make, Zapier or an AI agent

Actor ID: datagleaner/google-autocomplete-scraper

Minimal input:

{"keywords": ["coffee"], "sources": ["google"]}

Each tool below runs this Actor with your own Apify API token.

• n8n: add the Apify node (@apify/n8n-nodes-apify). On n8n Cloud you install it from the community node registry. Choose Run an Actor and get dataset, set Actor to datagleaner/google-autocomplete-scraper and paste the input above.

• Make: use the Apify app's Run an Actor module, then Get Dataset Items to read the results. Watch Actor Runs can trigger a scenario when a run finishes.

• Zapier: use the Apify action Run Actor, then the search Fetch dataset items. The trigger Finished Actor run starts a Zap when a run ends.

• AI agents (MCP): connect to https://mcp.apify.com/?tools=datagleaner/google-autocomplete-scraper. In Claude Code:

  claude mcp add --transport http apify "https://mcp.apify.com/?tools=datagleaner/google-autocomplete-scraper"
  

  Then run /mcp to sign in to Apify in your browser. Other clients can sign in with OAuth or send the header Authorization: Bearer YOUR_APIFY_TOKEN. Clients that run local MCP servers can use Apify's package (@apify/actors-mcp-server, run with npx -y and APIFY_TOKEN set) instead. Then ask the agent in plain words, for example:

  List the Google and YouTube autocomplete suggestions for these keywords: coffee, espresso machine, cold brew. Put them in a table with the source each came from.

• LangChain (Python):

# pip install langchain-apify, then set APIFY_TOKEN in your environment
import json
from langchain_apify import ApifyActorsTool
tool = ApifyActorsTool("datagleaner/google-autocomplete-scraper")
result = tool.invoke({"run_input": json.loads('{"keywords": ["coffee"], "sources": ["google"]}')})

FAQ

Do I need an API key or login? No. The Actor reads public, keyless endpoints.

Why did I get fewer results than my limit? Engines return a short list per query and many expansions overlap. Use alphabet or questions, or set recursionDepth to 2.

What happens on rate limits? HTTP 429 and server errors are retried with exponential backoff and the retry-after header. A query that still fails is skipped and the rest continue. Raise requestDelaySeconds for very large runs.

What happens if I start it with no keywords? It runs a small built-in example (5 Google suggestions for "coffee") so you can see the output, and says so in the run's status message. Add your own keywords for a real run.

Is the output in my language? Yes, set language and country. Seeds can be in any script.

Which sources can I combine? Any of Google, YouTube, Amazon and Bing in one run, all in the same row format with a source field.

Responsible use

This Actor collects public search suggestions only. Use the data in line with the terms of the engines you query and with applicable law. Suggestions can occasionally contain names of private persons; if you process personal data, the obligations are yours.