Google Autocomplete Keyword Scraper (+ YouTube & Bing)

Turn one seed keyword into hundreds of long-tail keyword ideas straight from the search boxes people actually type into. This Actor collects autocomplete suggestions (a.k.a. Google Suggest, "search suggestions", "People also search" style queries) from Google Search, YouTube, Google Shopping and Bing in bulk, expands each seed with A-Z, 0-9, question words, prepositions and comparisons, and returns one deduplicated row per keyword idea with the engines that suggested it, its best position and relevance score.

It's the free-tool workflow of AnswerThePublic / keywordtool.io / Ubersuggest's suggestion tab, automated for any number of seeds, any language and any country, with clean JSON/CSV/Excel output.

Why use this autocomplete scraper?

• 🔤 Multi-engine in one run: Google, YouTube, Google Shopping and Bing. A suggestion found on several engines appears once, with all its engines listed (engineCount is a great demand signal).
• 🌳 Expansion modes: a-z (seed a ... seed z), optional A-Z prefix, 0-9, questions (how, what, why, where, can, is...), prepositions (for, with, without, near, to...), comparisons (vs, alternative, or...), plus depth 2 (suggestions of suggestions).
• 🌍 Any market: set language (hl) and country (gl / Bing market), e.g. de + de, fr + ca, es + mx.
• 🧮 Ranking signals: bestPosition (1 = top suggestion), engine relevanceScore, engineCount, isQuestion, wordCount, containsKeyword.
• 🧹 Filters: must contain all seed words, negative keywords, cap per seed.
• ⚡ Fast and light: a 27-query A-Z expansion on Google takes a few seconds and runs on 256 MB.
• 🧾 A SUMMARY record with requests, failures and counts per seed.

Use cases

• SEO keyword research: long-tail and question keywords for blog posts, FAQ sections and programmatic SEO pages.
• YouTube SEO: video title and tag ideas from YouTube's own autocomplete.
• E-commerce & Amazon-style product research: buyer-intent modifiers from Google Shopping suggestions (brands, sizes, "for men", "sale").
• PPC: discover modifiers and negative keywords before launching Google Ads or Microsoft Ads campaigns.
• Content strategy & AI writing: feed real user questions into briefs, outlines or LLM prompts.
• Market & brand research: see what people associate with a brand or product ("brand vs", "brand alternative", "is brand legit").

How to use it

1. Enter one or more seed keywords.
2. Pick sources (Google, YouTube, Google Shopping, Bing) and expansions.
3. Set language and country, then click Start.
4. Export the dataset to CSV/Excel/Google Sheets, or call the Actor via the API, Zapier, Make or n8n.

Input example

{
  "keywords": ["coffee grinder", "best crm for"],
  "engines": ["google", "youtube", "bing"],
  "expansions": ["a-z", "questions"],
  "language": "en",
  "country": "us",
  "excludeWords": ["reddit"]
}

How many suggestions will I get?

Each request returns up to 15 suggestions (Google/YouTube/Shopping) or up to 25 (Bing in the US). Roughly, per seed keyword:

Settings	Requests per engine	Typical unique ideas per seed
No expansion	1	10–25
a-z	27	200–400
a-z + questions on 3 engines	40	600–1,200
a-z + depth 2	up to ~230	1,500–3,000

Use Max suggestions per seed or the run's Maximum cost per run to cap the output.

Output example

{
  "keyword": "coffee grinder",
  "suggestion": "coffee grinder for espresso",
  "engines": ["google", "youtube", "bing", "google-shopping"],
  "engineCount": 4,
  "bestPosition": 1,
  "relevanceScore": 1300,
  "expansionTypes": ["seed", "a-z"],
  "sourceQueries": ["coffee grinder", "coffee grinder f"],
  "depth": 1,
  "wordCount": 4,
  "isQuestion": false,
  "containsKeyword": true,
  "language": "en",
  "country": "us",
  "scrapedAt": "2026-10-19T15:47:11.311Z"
}

Rows are sorted per seed by depth, then number of engines, then best position, so the strongest ideas come first.

How much does it cost to scrape Google autocomplete?

You pay $0.0002 per unique keyword suggestion ($0.20 per 1,000). Duplicates across engines/expansions are merged before you pay, filtered-out suggestions are free, and there is no start fee.

Job	Typical suggestions	Cost
1 seed, Google, A-Z expansion (27 queries)	~200-250	~$0.05
1 seed, Google + YouTube + Bing, A-Z + questions	~500-700	~$0.10-0.14
1,000 keyword ideas	1,000	$0.20
50 seeds, Google, A-Z + questions	~15,000	~$3

• Apify's free plan includes $5 of monthly credit, i.e. about 25,000 keyword ideas a month for free.
• Paid Apify plans get automatic discounts: Silver -10%, Gold and above -20%.
• Use Max suggestions per seed or Maximum cost per run to cap spend.

Google autocomplete scraper vs. the alternatives

	This Actor	Typing into Google by hand	Free keyword tools (web UI)	Paid keyword suites
Google + YouTube + Google Shopping + Bing in one run	✅	❌	usually one engine	✅
A-Z, 0-9, questions, prepositions, "vs", depth 2	✅	❌	partial, daily caps	✅
Bulk seeds, any language/country	✅	❌	limited	✅
Deduplicated, with position/relevance/engine count	✅	❌	❌	✅
Search volume	❌ (use a volume tool on the output)	❌	sometimes	✅
Cost model	$0.20 per 1,000 ideas	free, slow	free with caps / upsell	monthly subscription

Use it as an API (Python, JavaScript, cURL, MCP)

Every Apify Actor is also a REST API, so you can call this Google autocomplete API from your own code, a no-code tool or an AI agent.

Python

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("datawright-labs/autocomplete-keyword-scraper").call(run_input={"keywords": ["coffee grinder"], "engines": ["google"], "expansions": ["a-z", "questions"]})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item)

JavaScript / Node.js

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: '<YOUR_APIFY_TOKEN>' });
const run = await client.actor('datawright-labs/autocomplete-keyword-scraper').call({"keywords": ["coffee grinder"], "engines": ["google"], "expansions": ["a-z", "questions"]});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items);

cURL (run and get results in one request)

curl -X POST "https://api.apify.com/v2/acts/datawright-labs~autocomplete-keyword-scraper/run-sync-get-dataset-items?token=<YOUR_APIFY_TOKEN>" \
  -H "Content-Type: application/json" \
  -d '{"keywords": ["coffee grinder"], "engines": ["google"], "expansions": ["a-z", "questions"]}'

AI agents (MCP)

Add https://mcp.apify.com/?tools=datawright-labs/autocomplete-keyword-scraper to Claude, Cursor or any MCP client and the agent can run it as a tool.

Results can also go straight to Google Sheets, Slack, Make, Zapier, n8n, Airbyte or a webhook via Apify integrations, and you can run it on a Schedule.

FAQ

Does it return search volume?

No. Autocomplete shows what people search for, ordered by popularity, but engines don't publish volumes. Use engineCount, bestPosition and relevanceScore as relative popularity signals, or pass the list to a search-volume tool.

Which endpoints does it use?

The public suggestion endpoints that browsers and search boxes use (Google's suggestqueries.google.com OpenSearch suggest service, Bing's OpenSearch suggest service). It sends a modest number of requests at a gentle rate, uses no login, and collects no personal data.

Some requests failed with HTTP 429/403. Very large runs from one IP can be rate-limited. The Actor retries automatically; for big runs enable Apify Proxy and/or lower Parallel requests. Failed requests are listed in the SUMMARY record and you never pay for them.

Can I get Amazon or DuckDuckGo suggestions?

Not at the moment. Open an issue if you need another source.

Is there a Google autocomplete API?

Google has no official, documented autocomplete API, but browsers use a public suggest endpoint. This Actor calls it (and YouTube, Shopping and Bing suggest) for you in bulk and returns clean, deduplicated rows over the Apify API.

How do I get YouTube search suggestions for video ideas?

Select the youtube engine. You get YouTube's own autocomplete for each seed and expansion, which is ideal for video titles and tags.

Is this an AnswerThePublic alternative?

For the "questions / prepositions / comparisons" part, yes: it produces the same style of long-tail and question keywords from live autocomplete, in bulk and in any language, without a monthly plan.

More tools from Datawright Labs

• Bulk PageSpeed & Core Web Vitals Checker: audit PageSpeed / Lighthouse scores for many URLs.
• Sitemap URL Extractor: get every URL of a site from its sitemaps.
• Website Screenshot API: bulk full-page screenshots and PDFs.
• Greenhouse, Lever & Ashby Jobs Scraper: normalized job feeds from 5 ATS platforms.
• Broken Link Checker: find broken links, images and redirects across a whole site.
• App Store Top Charts & Rank Tracker: daily top free/paid/grossing app rankings with rank changes.

Made by Datawright Labs: reliable, fairly priced web data tools. Questions or a missing feature? Open an issue on the Issues tab; we usually reply within a day.