Keyword Suggestions Scraper - Google, YouTube & Bing

Keyword Suggestions Scraper - Google, YouTube, Bing & Shopping Autocomplete

Turn a few seed keywords into hundreds of real search queries, straight from the autocomplete boxes of Google Search, YouTube, Google Shopping, Google News, Google Images and Bing, in one run. Most autocomplete scrapers cover one engine; this one queries six, expands each seed with a-z / 0-9, question words (who, what, how...) and prepositions (vs, for, with...), can go two levels deep for long-tail ideas, tags every question keyword, and has an only-new mode for monitoring autocomplete trends on a schedule. No start fee: you pay only per unique suggestion.

What data can you get

Field	Example
suggestion	coffee maker with grinder
engine	google, youtube, google-shopping, google-news, google-images, bing
seedKeyword	coffee maker
expansionQuery (the query that was typed)	coffee maker w
position (rank in the dropdown, 1 = top)	2
depth (1 = seed expansion, 2 = re-expanded suggestion)	1
country, language	us, en
wordCount	4
containsSeed	true
questionType	how (or null)
url (search-results page for the suggestion)	https://www.google.com/search?q=coffee%20maker%20with%20grinder&hl=en&gl=us
id, scrapedAt	google:coffee maker with grinder, 2026-11-06T12:25:59.836Z

Use cases

• SEO keyword research: long-tail keywords and content ideas people actually type into Google.
• YouTube video ideas and tags: what viewers search on YouTube around your topic.
• Ecommerce and Amazon-style product research: buyer-intent queries from Google Shopping and Bing ("air fryer deals", "... near me").
• FAQ and "People also ask" content: the questions expansion plus questionType gives you who/what/how/why questions ready for an FAQ page.
• Comparison content: the "vs" expansion finds competitor comparisons ("notion vs ...").
• Trend monitoring: schedule a run with Only new suggestions to get only autocomplete entries that appeared since last time.
• Local and international SEO: any country and language (gb/en, de/de, fr/fr...).

How to use

1. Enter one or more seed keywords.
2. Pick the engines.
3. Choose an expansion: none (1 query per seed and engine, ~10 suggestions), questions (19 queries), alphabet (37 queries) or full (55 queries).
4. Optionally set Depth 2, country and language, and Max suggestions (default 500, which caps the cost).

Quick check, two engines:

{
  "keywords": ["coffee maker"],
  "engines": ["google", "youtube"],
  "expansion": "none",
  "country": "us",
  "language": "en",
  "maxItems": 20
}

Full keyword research for SEO, Google and Bing:

{
  "keywords": ["standing desk", "ergonomic chair"],
  "engines": ["google", "bing"],
  "expansion": "full",
  "depth": 1,
  "country": "us",
  "language": "en",
  "maxItems": 1000
}

German market, questions only:

{
  "keywords": ["kaffeevollautomat"],
  "engines": ["google"],
  "expansion": "questions",
  "country": "de",
  "language": "de",
  "maxItems": 200
}

Weekly monitor of new YouTube suggestions (run on a schedule):

{
  "keywords": ["iphone 18"],
  "engines": ["youtube", "google"],
  "expansion": "alphabet",
  "onlyNew": true,
  "stateKey": "iphone-monitor",
  "maxItems": 500
}

Output example

Real items from a run of the quick-check input above:

[
  {
    "id": "google:coffee maker with grinder",
    "suggestion": "coffee maker with grinder",
    "engine": "google",
    "seedKeyword": "coffee maker",
    "expansionQuery": "coffee maker",
    "position": 2,
    "depth": 1,
    "country": "us",
    "language": "en",
    "wordCount": 4,
    "containsSeed": true,
    "questionType": null,
    "url": "https://www.google.com/search?q=coffee%20maker%20with%20grinder&hl=en&gl=us",
    "scrapedAt": "2026-11-06T12:25:59.836Z"
  },
  {
    "id": "youtube:coffee maker review",
    "suggestion": "coffee maker review",
    "engine": "youtube",
    "seedKeyword": "coffee maker",
    "expansionQuery": "coffee maker",
    "position": 2,
    "depth": 1,
    "country": "us",
    "language": "en",
    "wordCount": 3,
    "containsSeed": true,
    "questionType": null,
    "url": "https://www.youtube.com/results?search_query=coffee%20maker%20review",
    "scrapedAt": "2026-11-06T12:26:00.178Z"
  }
]

The run also writes a RUN_SUMMARY record (key-value store) with counts per seed and engine, request statistics and any errors.

Pricing

Pay per event, no start fee:

Event	Price
suggestion - one unique suggestion in the output	$0.0005

Examples: 1,000 suggestions = $0.50. A full expansion of one seed on one engine typically returns 150-250 unique suggestions (about $0.08-$0.13). Set Max suggestions or the run's "Maximum cost per run" to cap spending; the Actor stops cleanly when it is reached.

Tips and limits

• Each autocomplete query returns up to 10 suggestions (Bing up to ~15). Duplicates across queries are removed per engine, so full yields fewer than 55 x 10.
• The same suggestion on two engines is two rows (it tells you it ranks on both). Within one engine it is output once, with the first query that produced it.
• Max suggestions is shared fairly across engines: the engines are queried in turns, so with 3 engines and a limit of 300 each gets about 100. If one engine runs out of suggestions, the others use the rest.
• Depth 2 re-queries the top 10 suggestions of each seed and engine as plain queries (not re-expanded a-z), so it adds up to 10 queries per seed and engine.
• Suggestions are personalised by country and language only; no search volume or CPC data (autocomplete doesn't expose it).
• Google's shopping/news/images suggestions often overlap a lot with web search; YouTube and Bing differ the most.
• Rows from different engines are interleaved in the dataset; filter or sort by engine as needed.

FAQ

Is it legal? The Actor reads public autocomplete endpoints that any browser uses, without login. It collects search phrases, not personal data. You are responsible for using the data in line with the engines' terms and the laws that apply to you.

Do I need a proxy? No. If an engine rate-limits the run (HTTP 429), the Actor automatically retries through Apify's datacenter proxy. Residential proxies are not supported: if you select them, the Actor logs a warning and uses Apify datacenter proxy instead.

Why fewer suggestions than expected? Rare or very long seeds get few or no suggestions. Try a shorter seed or the alphabet expansion. Check RUN_SUMMARY for failed queries.

How does Only new suggestions work? The Actor remembers every suggestion it output per seed and engine in a named key-value store (keyword-suggestions-state, one record per Monitor name). Later runs skip those and charge only for new ones. Entries not seen for 180 days are forgotten.

Which languages work? Any two-letter Google hl / Bing language code, with proper accents and umlauts (e.g. "kühlschrank").

Integrations

Use the results via the Apify API (JSON, CSV, Excel), schedule runs, or connect with Make, Zapier, n8n, Google Sheets, Slack and webhooks. AI agents can call it through the Apify MCP server.

Changelog

• 0.1 (2026-11): First release. Google web, YouTube, Shopping, News, Images and Bing; alphabet / questions / full expansion; depth 2; question tagging; only-new mode.

Ready-to-run examples

• Find YouTube video ideas for sourdough bread
• Air fryer keywords from Google Shopping and Bing

More scrapers by the same developer

• Domain WHOIS Lookup - Bulk RDAP, Expiry, DNS & Availability
• AI Visibility Tracker: ChatGPT, Perplexity, Gemini, Claude
• Bing Search Scraper: Organic Results, Ads & Related Searches
• Google Ads Transparency Scraper
• Skool Scraper - Community Stats & Public Posts (No Login)
• Snapchat Scraper - Public Profiles, Spotlight & Story Stats

Browse all my Actors on my Apify Store profile.

Support

Found a bug or need another engine? Open an issue on the Actor's Issues tab.