Google Autocomplete Scraper: Google, YouTube, Amazon, Bing

Turn a few seed keywords into hundreds of real search suggestions. One actor covers Google Search, YouTube, Google Shopping, Amazon and Bing, in any language and country. Built for keyword research, content planning, PPC expansion and product research.

At a glance
You give it	Seed keywords, sources (Google, YouTube, Shopping, Amazon, Bing), expansion and depth
You get	One row per unique suggestion with its source, position, relevance and depth
Price	$0.005 per run + $0.00085 per suggestion (Free plan, lower on paid plans)
Speed	892 suggestions in about a minute
Needs	Nothing: no login, no API key, no proxy

What you get

• Five sources in one run: Google Search, YouTube, Google Shopping, Amazon (13 marketplaces) and Bing. Each shows a different side of demand: what people search, watch and buy.
• 15 Google suggestions per query with relevance scores, not the usual 10. The score is Google's own ranking signal.
• Expansion: alphabet (seed a ... seed z, seed 0 ... seed 9), questions (how seed, why seed, can seed...), and modifiers (seed for, seed vs, seed near, best...). Questions and modifiers use words in your language (English, German, French, Spanish, Italian, Portuguese, Dutch, Polish, Swedish, Japanese, Korean and Turkish), so café in French becomes pourquoi café and café pour, not how café, and コーヒー in Japanese becomes コーヒー とは and コーヒー おすすめ.
• Depth 1 to 3: every suggestion can be fed back in to find the suggestions behind it.
• Deduplicated per seed and source, with the exact query that produced each suggestion, its position, and its depth.
• Fast and light: one seed across Google, YouTube and Amazon with alphabet expansion returned 892 unique suggestions in under a minute, using about $0.005 of platform usage.

Use cases

• SEO keyword research. The long tail around a topic, straight from Google's predictions.
• Content ideas and FAQs. Question expansion gives you the questions people actually type.
• YouTube video ideas. What people search on YouTube, which is different from Google.
• Amazon product research and listing keywords. What shoppers type in the Amazon search bar, per marketplace.
• PPC keyword expansion. Real queries to add as keywords or negatives.
• International SEO. Same seed, different country and language.

Example inputs

One seed, three sources, alphabet expansion

{ "keywords": ["standing desk"], "sources": ["google", "youtube", "amazon"], "expand": ["alphabet"], "maxSuggestionsPerSeed": 300 }

Questions for a content plan

{ "keywords": ["intermittent fasting", "keto diet"], "sources": ["google"], "expand": ["questions"] }

Deep keyword tree

{ "keywords": ["project management software"], "expand": ["prepositions"], "depth": 2, "maxSuggestionsPerSeed": 1000 }

Amazon Germany

{ "keywords": ["kaffee"], "sources": ["amazon"], "language": "de", "country": "de", "expand": ["alphabet"] }

Input

Field	What it does
keywords	Seed keywords, one per line.
sources	Any of google, youtube, shopping, amazon, bing. Default google.
expand	Any of alphabet, questions, prepositions. Questions and prepositions are available in en, de, fr, es, it, pt, nl, pl, sv, ja, ko and tr; for other languages they are skipped with a warning, and alphabet still works.
depth	1 to 3. Each level queries the suggestions found at the level before. The seed stops at maxSuggestionsPerSeed, so if depth 1 already fills the cap there is no depth 2. The status message tells you when that happens.
maxSuggestionsPerSeed	Cap per seed and source. Default 300. The actor stops sending requests once it is reached. With depth 2 or 3, set it high (1,000+).
language, country	Two-letter codes. Amazon supports us, ca, mx, br, gb, de, fr, it, es, nl, jp, in, au.

Output

Real rows from a test run on 2026-09-26:

[
  {
    "suggestion": "standing desk converter",
    "seed": "standing desk",
    "source": "google",
    "query": "standing desk",
    "position": 1,
    "relevance": 601,
    "depth": 1,
    "wordCount": 3,
    "language": "en",
    "country": "us"
  },
  {
    "suggestion": "mat for standing desk anti-fatigue",
    "seed": "standing desk",
    "source": "amazon",
    "query": "standing desk a",
    "position": 1,
    "relevance": null,
    "depth": 1,
    "wordCount": 5,
    "language": "en",
    "country": "us"
  }
]

relevance is only available from Google Search. query is what was typed into the box; position is the suggestion's rank for that query.

Pricing

Pay per event:

• $0.005 per run start
• $0.00085 per unique suggestion on the Free plan (lower on higher Apify plans)

Apify platform usage is billed separately to your account and is tiny: 892 suggestions used about $0.005 (measured 2026-09-27).

Suggestions	Actor cost
15 (one seed, Google, no expansion)	about $0.018
300	about $0.26
1,000	about $0.86

Use maxSuggestionsPerSeed to control spend, or set a Maximum cost per run in the run options.

Automate it with n8n

Each workflow uses n8n's official Apify node, operation Run actor and get dataset, actor labrat011/google-autocomplete-scraper. Paste the input into Input JSON.

1. Content brief generator

Form Trigger (topic)
  > Apify: Run actor and get dataset   { "keywords": ["{{ $json.topic }}"], "expand": ["questions"], "sources": ["google"] }
  > Aggregate: all suggestions into one list
  > OpenAI / Anthropic: "Group these into H2 sections and write an outline"
  > Google Docs: create the brief

Switch Input JSON to Expression so the topic is filled in.

2. Weekly keyword tracker in Google Sheets

Schedule Trigger (Mondays)
  > Apify: Run actor and get dataset   (your 10 core keywords, alphabet expansion)
  > Google Sheets: Append rows with today's date
  > Compare with last week: new suggestions = emerging searches

3. YouTube video ideas to Notion

Schedule Trigger (weekly)
  > Apify: Run actor and get dataset   { "keywords": ["your niche"], "sources": ["youtube"], "expand": ["questions", "prepositions"] }
  > Filter: wordCount >= 4
  > Notion: create a page per idea

4. Amazon listing keywords

Manual Trigger
  > Apify: Run actor and get dataset   { "keywords": ["yoga mat"], "sources": ["amazon"], "expand": ["alphabet"], "depth": 2 }
  > Code: count how often each word appears across suggestions
  > Google Sheets: word frequency table for titles and backend keywords

5. PPC negative keyword finder

Manual Trigger
  > Apify: Run actor and get dataset   (your product keywords, prepositions expansion)
  > Filter: suggestion contains "free", "jobs", "diy", "used", "salary"
  > Google Sheets: candidate negatives for review

For AI agents

• Actor: labrat011/google-autocomplete-scraper
• Smallest input: { "keywords": ["protein powder"] }
• One row = one unique suggestion for one seed and source. Key fields: suggestion, seed, source, position, relevance, depth.
• Billing: apify-actor-start once per run, suggestion per row. Cap spend with maxSuggestionsPerSeed or a maximum cost per run.
• Run it: POST https://api.apify.com/v2/acts/labrat011~google-autocomplete-scraper/run-sync-get-dataset-items with the input as the JSON body, or call it from the Apify MCP server.
• Done signal: the run's status message starts with Saved N suggestions in R requests. It may add a note if the cap stopped a deeper level or an expansion was skipped for the language.

FAQ

Why does the same keyword give different suggestions on each source? Each engine predicts from its own search data. That is the point of running several.

How many suggestions does each query return? Google Search up to 15, YouTube, Shopping and Amazon up to 10, Bing usually 8 to 12 and sometimes more. The seed itself and anything already found for that seed are removed, so later queries add fewer new rows.

Which languages support questions and modifiers? English, German, French, Spanish, Italian, Portuguese, Dutch, Polish, Swedish, Japanese, Korean and Turkish. In Japanese, Korean and Turkish the question word goes after the seed, as people type it. For other languages use alphabet expansion, which works everywhere.

Do I need a proxy? No. Very large runs back off automatically if an engine slows down.

Support

Open an issue on the actor's Issues tab with the run ID and it will be looked at.