Google Autocomplete Scraper — Bulk Keyword Suggestions

Get the autocomplete suggestions that Google and YouTube show while people type, for many keywords at once, in any language and country. Turn one seed keyword into hundreds of real long-tail searches with A-Z, 0-9 and question expansion, deduplicated and ready for Excel, Google Sheets or your API.

What it's for

• SEO and keyword research: find the long-tail queries people actually type, by country and language.
• Content ideas: question expansion ("how", "why", "best", "vs", "near me"…) gives ready-made article and FAQ topics.
• PPC: discover new keywords and negative keywords for Google Ads campaigns.
• Product research: see which brands, models, sizes and features people search next to your product.
• YouTube: video ideas and titles from the YouTube search box.

How to use

1. Type one or more seed keywords (e.g. running shoes, crm software).
2. Choose the source (Google Search or YouTube), the language (en, it, de…) and the country (us, gb, it…). Suggestions really differ by country.
3. Pick the expansions:
   • A-Z: running shoes a, running shoes b … running shoes z
   • 0-9: running shoes 0 … running shoes 9
   • Questions and modifiers: built-in lists for English (how, what, why, where, when, who, which, can, is, best, vs, for, near me) and Italian (come, cosa, perché, dove, quando, chi, quale, migliore, vs, per, vicino a me). You can replace them with your own words in any language; write {keyword} to place a word after the keyword ({keyword} vs).
   • Depth 2: also queries every suggestion found at level 1. Much bigger output; capped by Max queries per keyword.
4. Run, then export the results as JSON, CSV or Excel, or read them via API.

Each suggestion is saved once per seed keyword, even if several queries return it. The seed keyword itself is not saved as a suggestion.

Input example

{
  "keywords": ["running shoes", "trail running"],
  "source": "google",
  "language": "en",
  "country": "us",
  "alphabet": true,
  "numbers": false,
  "questions": true,
  "maxDepth": 1
}

Output example

[
  {
    "seed": "running shoes",
    "query": "running shoes w",
    "suggestion": "running shoes with high heel drop",
    "position": 7,
    "expansionType": "alphabet",
    "language": "en",
    "country": "us",
    "source": "google"
  },
  {
    "seed": "running shoes",
    "query": "best running shoes",
    "suggestion": "best running shoes for men",
    "position": 2,
    "expansionType": "question",
    "language": "en",
    "country": "us",
    "source": "google"
  }
]

• query: the exact text sent to autocomplete
• position: rank of the suggestion in Google's list for that query (1-10)
• expansionType: base, alphabet, number, question or recursive (depth 2)

Pricing

Pay per result: $0.50 per 1,000 suggestions, charged only for unique suggestions saved to the dataset; the run stops automatically at the spending limit you set.

Notes

• Speed: about 350 unique suggestions in 15 seconds for one keyword with A-Z and questions, about 2,300 in 70 seconds at depth 2.
• The scraper sends few requests at a time and pauses between them. If Google answers "too many requests" (HTTP 429) it waits and retries; if the refusal continues, the run stops with a clear message instead of trying to bypass it: try again later or enable a proxy in the input.
• Suggestions are the public, non-personalised ones: they can differ slightly from what you see when signed in to Google.