Google Autocomplete Keyword Scraper: Web, YouTube, Images, News

What does Google Autocomplete Keyword Scraper do?

It collects the search suggestions Google shows as people type, for any seed keyword, country and language. Each record has the suggestion, the query that produced it, its position and relevance score, and the market it came from. Expansion options (a to z, question words, prepositions, recursive levels) turn one seed into hundreds of long tail keywords. Pricing is per result, and maxResults keeps runs small.

Data source

Google's public suggestion endpoint (suggestqueries.google.com), which returns live suggestions as JSON with no login.

Why use it?

• Build keyword lists and content plans from real search phrasing.
• Compare what people search in different countries and languages.
• Find YouTube, image, news and shopping phrasing, which differs from web search.
• Track how suggestions for a brand or product change over time with a schedule.

It replaces copying suggestions by hand from the search box.

How to use it

1. Open the Actor and enter seed keywords, country and language.
2. Click Start. Download results as JSON, CSV or Excel, or read them through the API.
3. Add a Schedule and an Integration (Slack, Google Sheets, webhook) to get new records automatically.

Input

Field	Description
seedKeywords	List of keywords to expand.
country	Two letter market code (gl), for example us, gb, de. Default us.
language	Two letter language code (hl), for example en, fr. Default en.
searchType	web, youtube, images, news or shopping. Default web.
expansion	none, alphabet, questions, prepositions or all. Default alphabet.
recursionDepth	0 to 3. Each level re-queries every new suggestion. Default 0.
maxResults	Hard stop for the run. You are billed per result, so this caps the cost.
proxyConfiguration	Apify proxy is recommended for larger runs.

Example input:

{
    "seedKeywords": ["car insurance", "home insurance"],
    "country": "us",
    "language": "en",
    "searchType": "web",
    "expansion": "alphabet",
    "recursionDepth": 0,
    "maxResults": 100
}

Output

Every record has the same shape:

{
    "id": "web:us:en:home insurance state farm",
    "suggestion": "home insurance state farm",
    "seedKeyword": "home insurance",
    "sourceQuery": "home insurance s",
    "position": 2,
    "relevance": 700,
    "suggestionType": "QUERY",
    "depth": 0,
    "expansion": "alphabet",
    "wordCount": 4,
    "country": "us",
    "language": "en",
    "searchType": "web",
    "scrapedAt": "2026-10-04T08:18:15.969Z"
}

Field	Meaning
id	Stable key (search type, country, language, suggestion), safe for deduplication between runs
suggestion	The suggested search phrase
seedKeyword	The seed it was found from
sourceQuery	The exact query typed to get it (seed plus letter, question word, or parent suggestion)
position	Rank in Google's list for that query
relevance	Google's relevance score when provided, otherwise null
suggestionType	Google's type label, usually QUERY, when provided
depth	0 for seed and expansions, 1 and up for recursive levels
expansion	seed, alphabet, question, preposition or recursive

Pricing

Pay per event: a small fee when a run starts, then a fixed price per result returned. You only pay for unique suggestions. Set maxResults to cap any run.

Limitations

• Google returns about 10 to 15 suggestions per query, so volume comes from expansion and recursion.
• Suggestions are not personalized, but they vary by time and market and may change between runs.
• Relevance and type are not returned for every search type.
• Search volume and related searches from result pages are not included.

FAQ

Is this legal? The Actor reads suggestions Google publishes for any visitor without a login and stores no personal data. You are responsible for how you use the data.

Something is missing or broken? Open an issue in the Issues tab; response time is usually under a day.