Google Autocomplete API: Keyword Suggestions & Questions Scraper

$4.00 per 1,000 expanded keywords — one keyword in, every Google Autocomplete suggestion out (typically 100-400 long-tail phrases per keyword with the a-z alphabet soup). Pay per result, no monthly rental. Keywords with no suggestion, errors and blocked requests are free.

This Actor turns seed keywords into long-tail keyword ideas from Google Search, YouTube or Google Shopping autocomplete, in any language and country. It queries the public suggest endpoint (the same one your browser uses), so there is no API key, no login and no browser: fast, cheap and stable.

What does it do?

For each seed keyword it sends:

• the keyword itself,
• alphabet soup / star expansion: keyword a … keyword z (Cyrillic or Greek letters for ru, uk, bg, el, or your own letters),
• optionally digits keyword 0 … keyword 9,
• optionally prefix letters a keyword … z keyword,
• optionally question words in the chosen language (how / what / why…, comment / pourquoi / quel…, wie / was / warum…, Spanish, Italian, Portuguese, Dutch).

All suggestions are merged, de-duplicated (case and spacing insensitive) and returned as one dataset item per keyword, with the seed query that produced each suggestion, its position and Google's relevance score.

How much does it cost?

What	Price
One expanded keyword (event keyword)	$0.004
1,000 expanded keywords	$4.00
Keyword with no suggestion, failed or blocked	$0

Example: a 250-keyword content plan with alphabet soup = 250 × $0.004 = $1.00, usually 40,000+ unique long-tail ideas. The price is the same whatever expansions you enable (digits, prefixes, questions). Proxy and compute are included.

Input

{
  "keywords": ["seo audit", "running shoes"],
  "language": "en",
  "country": "us",
  "source": "web",
  "expandAlphabet": true,
  "expandDigits": false,
  "expandPrefix": false,
  "expandQuestions": true,
  "onlyNew": false
}

Field	Description
keywords	Seed keywords (required). Duplicates are removed.
language	Google hl code: en, fr, de, es, it, pt, nl, ru, ja…
country	Google gl code: us, gb, fr, de, ca… Empty = no country bias.
source	web (Google Search), youtube or shopping.
expandAlphabet	keyword a…z (default on).
expandDigits	keyword 0…9.
expandPrefix	a…z keyword.
expandQuestions	Localized question words before the keyword.
customLetters	Your own characters instead of the default alphabet (e.g. abcdeéè, あいうえお).
onlyNew	Only suggestions not seen in previous runs (for schedules).
residentialFallback	Retry through residential proxies if Google rate-limits server IPs (default on, included in price).
maxConcurrency	Parallel queries per keyword (1-10, default 5).

Output

One item per keyword:

{
  "keyword": "seo audit",
  "language": "en",
  "country": "us",
  "source": "web",
  "suggestionsCount": 212,
  "suggestions": ["seo audit", "seo audit tool", "seo audit checklist", "seo audit free", "seo audit ai", "..."],
  "questions": ["how to do an seo audit", "what is an seo audit", "..."],
  "suggestionDetails": [
    {"suggestion": "seo audit tool", "seed": "seo audit", "seedType": "base", "position": 1, "relevance": 1250, "suggestionType": "QUERY", "seedCount": 3},
    {"suggestion": "seo audit ai", "seed": "seo audit a", "seedType": "suffix", "position": 1, "relevance": 1250, "suggestionType": "QUERY", "seedCount": 1}
  ],
  "suggestionsText": "seo audit\nseo audit tool\nseo audit checklist\n...",
  "onlyNew": false,
  "totalSuggestionsFound": 212,
  "newSuggestionsCount": null,
  "seedsQueried": 40,
  "seedsFailed": 0,
  "scrapedAt": "2026-10-16T08:00:00+00:00"
}

• seedType: base, suffix (a-z), digit, prefix, question.
• seedCount: how many seed queries returned this suggestion — a good proxy for how "central" the phrase is.
• suggestionsText: plain text, one suggestion per line, ready to paste into an LLM prompt or a RAG chunk.

Export as JSON, CSV, Excel or HTML from the dataset. Two ready-made views:

• Overview (?view=overview): one row per keyword with its suggestion and question lists.
• One row per suggestion (?view=suggestionList): a flat table (keyword, suggestion, seed, seedType, position, relevance, seedCount), ideal for Google Sheets or a keyword-clustering tool. CSV link: https://api.apify.com/v2/datasets/<datasetId>/items?view=suggestionList&format=csv.

Use cases

1. SEO content plan for a niche site

Who: an SEO consultant building topic clusters for a B2B site. Input:

{"keywords": ["account based marketing", "abm software", "intent data"], "language": "en", "country": "us", "expandAlphabet": true, "expandQuestions": true}

Output (excerpt): ["account based marketing examples", "account based marketing vs demand gen", "what is account based marketing", "abm software for small business", ...] — about 600 phrases for $0.012. Next: paste suggestionsText into an LLM to cluster phrases into pillar pages and FAQs, or export CSV to a keyword tool for volumes.

2. Weekly trend watch in Google Sheets

Who: a French e-commerce marketer selling e-bikes. Input (scheduled every Monday):

{"keywords": ["vélo électrique", "trottinette électrique"], "language": "fr", "country": "fr", "onlyNew": true}

Output (excerpt): {"keyword": "vélo électrique", "newSuggestionsCount": 7, "suggestions": ["vélo électrique prime 2027", "vélo électrique pliant léger", ...]} Next: the Apify Google Sheets integration appends only the new phrases; the team writes product pages before competitors. Cost: 2 × $0.004 per week, and $0 for a week with nothing new.

3. YouTube video ideas for a creator agency

Who: a content strategist planning 50 videos per month. Input:

{"keywords": ["notion tutorial", "obsidian", "second brain"], "source": "youtube", "language": "en", "country": "gb", "expandQuestions": true}

Output (excerpt): ["notion tutorial for students", "notion tutorial 2026", "how to use obsidian for beginners", ...] Next: n8n sends the list to an LLM that drafts titles, then to a Trello board.

4. Marketplace listing keywords

Who: an Amazon/Shopify seller writing product titles. Input: {"keywords": ["yoga mat", "resistance bands"], "source": "shopping", "expandDigits": true} Next: use the most frequent seedCount phrases in titles and bullet points.

API examples

Python

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("<USERNAME>/google-autocomplete").call(run_input={
    "keywords": ["seo audit", "running shoes"], "language": "en", "country": "us",
})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["keyword"], item["suggestionsCount"], item["suggestions"][:5])

JavaScript

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: '<YOUR_APIFY_TOKEN>' });
const run = await client.actor('<USERNAME>/google-autocomplete').call({
    keywords: ['seo audit', 'running shoes'], language: 'en', country: 'us',
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
items.forEach((i) => console.log(i.keyword, i.suggestionsCount));

HTTP (one call, returns items)

POST https://api.apify.com/v2/acts/<USERNAME>~google-autocomplete/run-sync-get-dataset-items?token=<TOKEN>
{"keywords": ["seo audit"], "language": "en", "country": "us"}

Integrations (Make, Zapier, n8n, Google Sheets)

• Make / Zapier: use the Apify "Run Actor" module with the input above, then "Get dataset items" → Google Sheets "Add row" with keyword and suggestionsText.
• n8n: HTTP Request node on the run-sync-get-dataset-items URL → Split Out on suggestions → your LLM or database node.
• Google Sheets: Apify's native integration exports each run to a sheet; combine with a schedule and onlyNew: true for a weekly "new keywords" tab.
• Webhooks: get notified when a run finishes.

Use with AI agents (MCP)

This Actor is available through the Apify MCP server, so Claude, Cursor, ChatGPT and other AI agents can call it as a tool: "find long-tail keywords and questions for X in French". The output is compact JSON plus suggestionsText (plain text), so agents can read it without extra parsing. Typical agent flow: call with 3-10 keywords → cluster suggestions → draft an outline per cluster.

FAQ

Do I need a Google API key or account? No. The Actor uses Google's public autocomplete endpoint. No login, no cookies.

Why one charge per keyword and not per suggestion? You pay for the work you ask for: one seed keyword fully expanded = $0.004, whether it yields 50 or 400 suggestions.

Is search volume included? No. Google Autocomplete doesn't publish volumes. The relevance score and seedCount give a relative ranking; use a keyword planner for absolute volumes.

Which languages work? Any language Google supports. Question words are built in for en, fr, de, es, it, pt, nl (English is used for others). Cyrillic and Greek alphabets are automatic; use customLetters for others (e.g. Japanese kana).

What happens when my budget (max charge) is reached? The Actor stops cleanly; it never delivers unpaid results.

What about rate limits? The Actor spreads queries, retries, and switches to proxies when Google limits an IP. Failed queries are reported in seedsFailed; a keyword where every query failed is not charged.

Limitations

• Google returns at most ~10-15 suggestions per query, so a keyword with only alphabet soup yields at most ~400 phrases (usually fewer after de-duplication).
• Suggestions are personalised and vary slightly over time and by location; results reflect a neutral, logged-out user.
• No search volume, CPC or difficulty.
• shopping source uses Google's shopping autocomplete; if Google returns web suggestions instead, they are still delivered.
• Very rare or nonsense keywords return nothing (free).