Keyword Research Tool · Google, YouTube, Amazon, Bing, eBay & DuckDuckGo

Turn a handful of seed terms into thousands of real, long-tail keyword ideas — pulled live from the autocomplete of six major search engines, deduplicated, and ranked by cross-engine popularity. No API keys, no monthly seat fees. Pay only for the keywords you get.

This is the pay-as-you-go, bulk, multi-engine alternative to tools like KeywordTool.io and AnswerThePublic — built for SEO teams, content creators, Amazon/eBay sellers, PPC managers, and automated content pipelines (n8n, Make, Zapier, MCP, AI agents).

Why this actor

• 6 engines in one run — Google, YouTube, Amazon, Bing, eBay, DuckDuckGo. Web SEO intent, video intent, and e-commerce buyer intent, side by side.
• Real autocomplete data — the exact "what people actually type" suggestions the engines serve, not made-up permutations.
• Massive expansion — each seed is grown with question words (who/what/how…), prepositions & modifiers (for/vs/best/cheap…), and A–Z / 0–9 variants to surface hundreds of long-tails per seed.
• One clean row per keyword — deduplicated across engines, with a frequency score (how many engine/query forms produced it), the best autocomplete rank, and Google's relevance signal.
• Question detection — instantly filter the question keywords that make great blog/FAQ/briefs.
• Multi-country & multi-language — localized suggestions per country (Amazon/eBay auto-map to the right marketplace).
• Monitoring mode — schedule runs and return only newly-appearing keywords via a persistent dedup store. Track emerging long-tails and trends over time.
• No cookies, no login, no captchas. Fast lightweight HTTP.

Use cases

You are a…	You use it to…
SEO / content marketer	Build topic clusters, find question keywords for briefs, expand a niche
YouTuber / video team	Discover what viewers search on YouTube for your topic
Amazon / eBay seller	Find buyer-intent product keywords & listing terms per marketplace
PPC manager	Mine long-tail positive & negative keywords across engines
Agency / SaaS builder	Power a keyword feature or content pipeline via the Apify API/MCP

Input

Field	Type	Description
keywords	array	Required. Seed terms to expand (e.g. coffee maker, crm software).
engines	array	Any of google, youtube, amazon, bing, ebay, duckduckgo.
modifiers	array	questions, prepositions, alphabet, alphabet_prefix, numbers.
includeSeed	boolean	Also query the bare seed's own suggestions. Default true.
country	string	Two-letter country code (e.g. us, gb, de, in).
language	string	Two-letter language code (e.g. en, de, fr).
matchType	string	any, or seed_included (keyword must contain the seed).
minLength	integer	Drop suggestions shorter than N characters.
maxKeywordsPerSeed	integer	Cap keywords per seed (keeps the highest-ranked). 0 = no cap.
dedupAcrossRuns	boolean	Return only NEW keywords vs previous runs (monitoring).
dedupStoreName	string	Named store per watchlist for monitoring.
concurrency	integer	Parallel autocomplete requests (1–20).
proxyConfiguration	object	Apify Proxy recommended for geo-consistency.

Example input

{
  "keywords": ["coffee maker", "cold brew"],
  "engines": ["google", "youtube", "amazon"],
  "modifiers": ["questions", "prepositions"],
  "country": "us",
  "language": "en"
}

Output

Each unique keyword is one dataset item:

{
  "keyword": "which coffee maker is best",
  "seed": "coffee maker",
  "sources": ["google", "youtube"],
  "sourceCount": 2,
  "frequency": 5,
  "bestRank": 0,
  "relevance": 951,
  "isQuestion": true,
  "isSeed": false,
  "wordCount": 5,
  "characterCount": 26,
  "modifiers": ["questions"],
  "country": "us",
  "language": "en"
}

• frequency — how many engine/query forms surfaced this keyword. Higher = more broadly suggested (a strong relative popularity proxy).
• bestRank — best (lowest) position it held in any autocomplete list. 0 = top suggestion.
• relevance — Google's own relevance score when the keyword came from Google (higher = stronger).
• sources / sourceCount — which engines suggested it; keywords found by several engines are usually the most valuable.

A run summary (seeds, engines, calls made, unique keywords, keywords returned, skipped-as-seen) is saved to the key-value store under SUMMARY.

Pricing

Pay-per-event: you're charged per unique keyword returned (plus a small per-run start fee). No monthly subscription, no seats. A typical single-seed run with the questions modifier across 3 engines returns a few hundred keywords. See the Pricing tab for the live per-event rate.

Tip: use maxKeywordsPerSeed and matchType: seed_included to keep runs tight and costs predictable.

FAQ

Does this give exact monthly search volume? No — Google/Amazon don't expose true volume without a paid Ads account. Instead you get a robust relative popularity signal (frequency, bestRank, Google relevance) that's ideal for prioritizing long-tails. This is the same autocomplete data that powers popular paid keyword tools.

How many keywords will I get? Depends on seeds × modifiers × engines. One seed with questions across 3 engines is typically ~300–600 unique keywords. Add alphabet for thousands.

Which engines for which job? Google/Bing/DuckDuckGo = web SEO; YouTube = video; Amazon/eBay = e-commerce buyer intent.

Can I monitor for new keywords? Yes — set dedupAcrossRuns: true with a dedupStoreName and schedule it. Each run returns only keywords not seen before.

Limitations

• Autocomplete returns up to ~10–15 suggestions per query form; expansion (modifiers) is what scales coverage.
• Relevance scores are only provided by Google.
• Some engines geo-target by IP; use Apify Proxy with a matching country for best localization.