Google Autocomplete Keyword Tool + Content Gap Finder (10 engines)

Turn one seed keyword into hundreds of real long-tail keywords, the searches people actually type, from the autocomplete of ten engines at once:

• Google;
• YouTube;
• Bing;
• Amazon;
• eBay;
• DuckDuckGo;
• Yahoo;
• Google Shopping;
• TikTok;
• Baidu (百度).

Every keyword is ranked by how many engines suggest it. Give it your website, and it also tells you which keywords you have no page for yet: your content gap.

Use it for SEO keyword research, blog and video topic planning, Amazon and eBay product listings, YouTube and TikTok titles, PPC keyword lists, FAQ pages, and AI agents that need fresh search demand.

• 🔎 Ten autocomplete engines in one run. Pick any mix: web search, video (YouTube, TikTok), shopping (Amazon, eBay, Google Shopping) and China (Baidu).
• 📊 A free demand signal. Autocomplete has no search volume, but a keyword that Google, Bing and YouTube all suggest is one many people type. engineCount and bestPosition rank the list, and the strongest keywords come first.
• 🕳️ Content gap. Paste your site or sitemap URL. Every keyword then gets covered, coverage and the closestPage on your site, so the uncovered ones are your next articles.
• 🧠 Optional AI topics and content ideas. Keywords are grouped into topics, and each gets a title for the page or video that would answer it.
• ❓ Expansions that find the long tail:
  • alphabet (seed a … seed z) and numbers;
  • question words (how, what, why … in English, German, French, Spanish, Italian, Portuguese and Dutch);
  • prepositions (for, with, vs, near …);
  • buying words (best, cheap, review, price, alternative …).
• 🌍 Any country. Google, Bing, DuckDuckGo, Amazon and eBay results are localised (us, gb, de, fr, es, it, ca, au, in, jp …), and requests come from a residential IP in that country.
• ✅ No login, no API key, no proxy setup.

Which engines does it cover?

Engine	Best for
Google	SEO, blog topics, the biggest search volume
YouTube	video titles, tutorials, channel ideas
Bing	a second web-search opinion
DuckDuckGo, Yahoo	more web-search agreement for the engineCount signal
Amazon	product research, listing keywords, buyer phrases
eBay	resale and second-hand demand, sizes and variants
Google Shopping	shopping-intent phrases
TikTok	what people search on TikTok: trends, how-tos, products
Baidu (百度)	Chinese search demand (use Chinese seeds)

How much does it cost?

You pay per keyword delivered, and there's no monthly fee. A keyword found from several seeds or engines is delivered and charged once.

Event	Price (FREE/BRONZE)	Lower on paid plans
Keyword (engines, positions, intent, content gap)	$0.001 per keyword ($1 per 1,000)	down to $0.0008
Keyword with AI topic and content idea	$0.003 per keyword ($3 per 1,000)	down to $0.0025
Actor start	$0.00005 per run	—

Examples:

• 200 keywords for one seed cost $0.20.
• The same 200 with AI topics and content ideas cost $0.60.

maxKeywordsPerSeed caps what you pay, and so does Apify's maximum cost per run. The actor stops before it would go over.

AI topics are a paid-plan feature. Free-plan runs include a 3-keyword trial, and the rest of the keywords are saved without the AI fields at the plain keyword price.

How fast is it?

Measured with the default settings (alphabet + questions on Google, YouTube, Bing and Amazon, so 160 requests per seed). One seed took about 10 seconds and found 660–1,220 unique keywords ("crm software" 662, "sourdough bread" 1,052, "running shoes" 1,224). On the platform, the example input (50 keywords) finished in 11 seconds on 256 MB. A run with AI and a content gap (40 keywords, 612 site pages) took 30 seconds.

What data do you get?

One record per keyword. This is real output from a run with siteUrl: https://backlinko.com and AI on:

{
  "keyword": "what is link building",
  "seed": "link building",
  "engines": ["bing", "google", "youtube"],
  "engineCount": 3,
  "positions": {"bing": 4, "google": 1, "youtube": 1},
  "bestPosition": 1,
  "foundVia": "what link building",
  "expansion": "questions",
  "wordCount": 4,
  "isQuestion": true,
  "intent": "informational",
  "covered": true,
  "coverage": 1,
  "closestPage": "https://backlinko.com/link-building",
  "topic": "Link Building Basics",
  "contentIdea": "What Is Link Building: A Comprehensive Guide for Beginners",
  "country": "us",
  "language": "en",
  "scrapedAt": "2026-10-13T23:47:05+00:00"
}

And one from the same run that the site does not cover yet, so it's a content gap:

{"keyword": "what is link building in digital marketing", "engines": ["bing", "google"], "engineCount": 2,
 "bestPosition": 3, "intent": "informational", "covered": false, "coverage": 0.5,
 "closestPage": "https://backlinko.com/link-building", "topic": "Link Building Basics",
 "contentIdea": "The Role of Link Building in Digital Marketing"}

Field	Meaning
engineCount, engines, positions, bestPosition	Which engines suggested it and how high (1 = first suggestion). More engines plus a higher position means stronger demand.
foundVia, expansion	The query that first returned it, and which expansion produced it (seed, alphabet, numbers, questions, prepositions, commercial, deeper).
intent	informational, commercial (best, review, vs …), transactional (buy, price, near me …) or navigational (login, app …), from rules on the words.
isQuestion, wordCount	For FAQ pages and long-tail filters.
covered, coverage, closestPage	Content gap fields (need siteUrl, otherwise null).
topic, contentIdea	AI fields (need aiInsights, otherwise null).

You can download the dataset as JSON, CSV, Excel or HTML, or read it through the API.

Content gap: find the keywords your site has no page for

Set siteUrl to your site (https://example.com), to one section (https://example.com/blog), or straight to a sitemap (https://example.com/sitemap.xml).

• The actor reads your sitemaps: for a section such as /blog, its own sitemap_index.xml, wp-sitemap.xml and sitemap.xml first; then those listed in robots.txt, or else /sitemap.xml, /sitemap_index.xml and /wp-sitemap.xml. Sitemap indexes and .xml.gz files are followed.
• For each keyword it finds the page whose address contains most of the keyword's words.
• coverage is the share of the keyword's meaningful words found in that page's address. Small words like "how", "the" and "for" don't count, and plurals match singulars.
• A keyword is covered at 0.75 or more.

Sort by covered = false and engineCount to get your content plan.

How it matches, honestly: it compares words in your page addresses (URL slugs), not page text. That makes it fast and free of extra page downloads, and it works well on sites with descriptive URLs such as /blog/how-to-clean-running-shoes.

• It is word overlap, not meaning, so word order is ignored: "who is building link" (a question about a property-management app) matches a /link-building page.
• Pages with ID-only URLs (/p/12345) can't be matched.
• Keywords in Chinese, Japanese or Korean get coverage: null.

AI topics and content ideas

Turn on aiInsights. The strongest keywords of each seed are sent to Google's Gemini in batches, and each gets:

• topic: a 2–4 word topic label. Related keywords share the exact same label, so you can group by it and treat each topic as a cluster or pillar page.
• contentIdea: a title for the page or video that would answer that search, with an angle such as a guide, checklist, comparison or mistakes to avoid. The AI is told never to invent years, statistics, brands or places that the keyword doesn't contain.

Ranks, engines, positions, intent and coverage are always computed by code, never by the AI. If the AI service is down, the run keeps going, and the keywords are saved and charged as plain keywords.

How to use it

1. Enter one or more seed keywords, for example running shoes.
2. Pick the engines. The default is Google, YouTube, Bing and Amazon.
3. Pick the expansions. The default is alphabet + questions; add prepositions and buying words for more English long tail.
4. Optional:
   • set country and language;
   • add your siteUrl for the content gap;
   • turn on aiInsights;
   • turn on deeper to re-expand the 25 strongest suggestions.
5. Run it and export the keywords.

Input example

{
  "keywords": ["running shoes", "trail running"],
  "engines": ["google", "youtube", "bing", "amazon"],
  "expansions": ["alphabet", "questions", "prepositions"],
  "maxKeywordsPerSeed": 300,
  "country": "us",
  "language": "en",
  "siteUrl": "https://example.com/blog",
  "aiInsights": true
}

Ready-made keyword tasks

Open one, adjust it, and run or schedule it:

• YouTube Keyword Suggestions (Autocomplete) Tool
• Amazon Keyword Research from Autocomplete
• Content Gap Finder: Keywords Your Site Misses
• Question Keywords for Blog Posts
• TikTok Search Suggestions Keyword Tool

Use it with AI agents (MCP)

https://mcp.apify.com?tools=bossless/autocomplete-keyword-expander

Ask your assistant "find long-tail keywords for 'standing desk' on Google and Amazon" or "which 'link building' questions does backlinko.com not cover yet?". It runs this actor and reads back ranked keywords. Keep agent runs small (maxKeywordsPerSeed: 50, one expansion) so they finish in seconds.

Use the API from code

curl -X POST "https://api.apify.com/v2/acts/bossless~autocomplete-keyword-expander/run-sync-get-dataset-items?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"keywords": ["standing desk"], "engines": ["google", "amazon"], "expansions": ["questions"], "maxKeywordsPerSeed": 50}'

from apify_client import ApifyClient
client = ApifyClient("YOUR_TOKEN")
run = client.actor("bossless/autocomplete-keyword-expander").call(run_input={
    "keywords": ["sourdough bread"], "engines": ["google", "youtube", "bing"], "siteUrl": "https://example.com"})
for k in client.dataset(run["defaultDatasetId"]).iterate_items():
    if k["covered"] is False:
        print(k["engineCount"], k["keyword"])

FAQ

Does it give search volume? No. Autocomplete doesn't publish volumes, and real volume data comes from paid keyword databases. What you get instead is free and current:

• engineCount: how many engines suggest the keyword;
• bestPosition: how high they put it.

Engines only suggest what many people type, so a keyword suggested by three engines in first or second place is a strong one.

How many keywords will I get? It depends on the seed and the settings. Our tests with the defaults (alphabet + questions on four engines) found 660–1,220 unique keywords per seed. maxKeywordsPerSeed keeps the strongest N, 200 by default.

Can I research other countries and languages? Yes. Set country (two letters, such as de) and language (such as de). Google, Bing, DuckDuckGo, Amazon (13 marketplaces) and eBay (8 sites) are localised, and question words exist for seven languages. Prepositions and buying words are English only. Use Chinese seeds with Baidu.

What if an engine blocks a request? It is retried on new IP addresses, and the other engines still deliver. A run fails only if every request was blocked and nothing was delivered. A seed that no engine has suggestions for gives an empty, successful run.

What if my site has no sitemap? The keywords are still delivered, with the content-gap fields left empty, and the run's status message says so. Give the sitemap URL directly if it lives somewhere unusual.

Is this an official Google, YouTube, Amazon, Bing, TikTok or Baidu product? No. This is an independent tool and is not affiliated with or endorsed by any of these companies.

Is it legal? It reads the public suggestions each engine shows in its search box, one query at a time. You are responsible for how you use the data.

Related tools

• China Hot Search in English: what Weibo, Douyin and Baidu users are searching right now, translated.
• Zhihu Scraper (知乎): questions, answers, hot list and search suggestions from China's Q&A site.

---

关键词挖掘工具（百度下拉词 / 谷歌下拉词）：从 Google、YouTube、Bing、Amazon、TikTok、百度等 10 个搜索引擎的下拉联想词中批量挖掘长尾关键词，按出现的引擎数排序，并可对照你的网站地图找出尚未覆盖的关键词（内容缺口）。无需登录。