Keyword Research Tool - Google, YouTube, TikTok & Amazon avatar

Keyword Research Tool - Google, YouTube, TikTok & Amazon

Pricing

from $0.38 / 1,000 keywords

Go to Apify Store
Keyword Research Tool - Google, YouTube, TikTok & Amazon

Keyword Research Tool - Google, YouTube, TikTok & Amazon

Keyword research from real autocomplete: one seed becomes hundreds of long-tail keywords and questions across Google, YouTube, TikTok, Bing, Amazon, Walmart, eBay, DuckDuckGo, Yandex, Naver & Wikipedia, each with search intent, rank and cross-platform demand. Monitor mode returns only new keywords.

Pricing

from $0.38 / 1,000 keywords

Rating

5.0

(1)

Developer

Scrape Sage

Scrape Sage

Maintained by Community

Actor stats

0

Bookmarked

214

Total users

36

Monthly active users

15 days ago

Last modified

Features

Limited permissions

Share

Disclaimer: This Actor is an independent tool and is not affiliated with, endorsed by, or sponsored by Google, YouTube, TikTok, Amazon, Walmart and the other search engines and sites it queries, or any of their owners or subsidiaries. All trademarks mentioned are the property of their respective owners. Site names are referenced only to describe the publicly available sources this Actor collects data from.

Turn one seed keyword into hundreds of real long-tail keywords that actual people are searching for - pulled live from the autocomplete of Google, YouTube, TikTok, Bing, Amazon, Walmart, eBay, DuckDuckGo, Yandex, Naver, Wikipedia, Google Images, Google News and Google Shopping. Every keyword is enriched with search-intent classification, SEO metrics and a cross-platform demand signal, so you get a research-ready dataset, not just a list of strings.

No login, no API key, no browser - fast JSON reads straight from the suggest endpoints each search box uses.

Why this keyword research tool?

This actor queries 14 search boxes in one run - Google, YouTube, TikTok, Bing, Amazon, Walmart, eBay, DuckDuckGo, Yandex, Naver, Wikipedia, Google Images, Google News and Google Shopping - and merges the results, so you instantly see which keywords appear on multiple platforms (the strongest demand signal there is).

FeatureThis actor
Google autocomplete✅
YouTube, Bing, DuckDuckGo✅ one run
TikTok search suggestions (TikTok SEO, video and creator content)✅ one run
Amazon, Walmart & eBay (product / marketplace keywords)✅ one run
Yandex and Naver (Russian, Turkish and Korean markets)✅
Wikipedia, Google Images / News / Shopping✅
Alphabet · question · preposition · comparison · commercial expansion✅ all
Recursive deep expansion✅
Search-intent classification (informational / commercial / transactional / navigational)✅
SEO metrics (word count, long-tail flag, contains-seed)✅
Google relevance score per suggestion✅
Cross-platform demand (platformCount + full sources)✅
Entity data - Wikipedia description, Wikidata id, page URL & thumbnail; YouTube Knowledge Graph entities✅
eBay category matches for each eBay query✅
Countries and languages by code or name (DE, Germany, es, Spanish)✅
Added-words breakdown (the modifier each keyword adds to its seed)✅
Monitor mode - only new keywords✅

Platforms covered

PlatformWhat it tells you
Google (web)Core SEO & content keyword demand
YouTubeVideo SEO - what viewers search for, with the Knowledge Graph entity (name, description, image, link) where YouTube shows one
TikTokTikTok SEO - what people type into TikTok's search bar, in the region and language you choose (suggestions TikTok tags with another language are left out)
BingSecond search engine, often different long-tail
DuckDuckGoPrivacy-audience search behaviour
YandexWhat Yandex users type: strongest for Russian, Turkish and other CIS-market keywords (an English seed can also return Russian phrases, such as "... перевод", a translation search)
NaverKorea's largest search engine, for Korean keywords (read when the country is KR)
AmazonProduct & buyer keywords (Amazon SEO / PPC), read in the Amazon store of your country (23 stores, from the US, UK and Germany to Japan, India, Brazil and the UAE)
WalmartProduct keywords shoppers type on walmart.com, for Walmart Marketplace listings and ads (the US store, read when the country is US)
eBayMarketplace & resale product keywords in the eBay site of your country (14 countries), plus the eBay categories (id + name) each query falls under
WikipediaEntities, topics and disambiguation terms - each with its short description, Wikidata id, page id, URL and thumbnail
Google ImagesVisual-intent queries
Google NewsTrending / newsworthy angles on a topic
Google ShoppingHigh-commercial-intent product queries

Use cases

  • SEO & content strategy - discover the long-tail and question keywords to build topic clusters, blog briefs and FAQ / People-Also-Ask sections.
  • PPC & Amazon/Walmart/eBay ads - find buyer-intent keywords (transactional / commercial) and negative keywords for Google Ads and marketplace campaigns.
  • TikTok SEO - find the phrases people search on TikTok for your topic, to use in captions, on-screen text and video titles.
  • App Store Optimization (ASO) - pair this with the Google Play / App Store scrapers to research the terms users search for in stores.
  • E-commerce product research - Amazon, Walmart & eBay autocomplete reveals what shoppers actually type, including brands, sizes and modifiers.
  • YouTube & video - surface video titles and topics your audience is already searching for.
  • Trend monitoring - schedule the actor in monitor mode to capture only newly emerging long-tail keywords for a topic over time.
  • Market & competitor research - see which keywords show demand across multiple platforms, and in any country/language.

How to use

  1. Sign up for Apify - the free plan is enough to try this actor.
  2. Open the Keyword Research Tool, enter your seed keywords, pick the platforms and expansions, and click Start.
  3. When the run finishes, your keywords are in the dataset table on the Output tab, and the run's status message says what was found.
  4. Export as JSON, CSV, Excel, XML, or RSS - or pull results programmatically via the Apify API.

Input

{
"keywords": ["coffee maker", "crm software"],
"platforms": ["google", "youtube", "bing", "amazon"],
"expansions": ["alphabet", "questions", "prepositions"],
"country": "US",
"language": "en",
"recursiveDepth": 0,
"maxResults": 1000,
"dedupeAcrossPlatforms": true,
"includeQuestionsOnly": false,
"monitorMode": false
}
  • keywords (required) - one or more seed keywords; each is expanded across every selected platform.
  • platforms - any of google, youtube, bing, amazon, ebay, duckduckgo, wikipedia, google_images, google_news, google_shopping, tiktok, walmart, yandex, naver.
  • expansions - alphabet (a–z), numbers (0–9), questions (how/what/why…), prepositions (for/with/near…), comparisons (vs/alternative…), commercial (best/buy/price…). Empty = only the bare seed.
  • country / language - localise suggestions (Google gl/hl, Amazon/eBay store, Bing market, Wikipedia edition, TikTok region and language, Yandex language; Walmart is read for US and Naver for KR). Type a code in any case (US, uk, de) or a name (Germany, Spanish); a regional language such as pt-BR works too.
  • recursiveDepth (default 0) - re-query discovered keywords to dig deeper into the long tail (bounded by maxResults).
  • maxResults (default 1000) - hard cap on the keywords delivered (cost control). Only keywords that pass your filters count, so a filtered run still fills it, and several seeds share it side by side.
  • dedupeAcrossPlatforms (default true) - one row per keyword with a platforms list + sources array, or one row per keyword-per-platform when off.
  • includeQuestionsOnly / minWordCount / mustIncludeTerms / excludeTerms - filters, applied before maxResults counts. Terms match at the start of a word, case-insensitive (review also keeps reviews; tea does not match steam).
  • monitorMode (default false) - return ONLY keywords not seen in previous runs (see below).
  • proxyConfiguration - with the default, each request first tries a direct connection (the fastest way to reach these autocomplete endpoints) and falls back to the Apify proxy on any miss; choose a proxy group or your own proxies to route every request exactly as you set it.

Output

One record per unique keyword (or per keyword-per-platform when dedupeAcrossPlatforms is off):

{
"keyword": "coffee maker alarm clock",
"seed": "coffee maker",
"seeds": ["coffee maker"],
"platform": "google",
"platforms": ["google", "youtube", "bing"],
"platformLabels": ["Google", "YouTube", "Bing"],
"platformCount": 3,
"country": "US",
"language": "en",
"searchIntent": "general",
"isQuestion": false,
"isCommercial": false,
"isTransactional": false,
"isComparison": false,
"isLocal": false,
"wordCount": 4,
"charCount": 24,
"isLongTail": true,
"containsSeed": true,
"bestRank": 3,
"bestRelevanceScore": 603,
"expansionTypes": ["alphabet"],
"discoveryDepth": 0,
"entity": null,
"sources": [
{ "platform": "google", "platformLabel": "Google", "rank": 3, "relevanceScore": 603, "suggestionType": "QUERY", "subtypes": [512], "modifierQuery": "coffee maker a", "expansionType": "alphabet", "depth": 0 },
{ "platform": "youtube", "platformLabel": "YouTube", "rank": 4, "relevanceScore": null, "suggestionType": null, "subtypes": [512], "modifierQuery": "coffee maker a", "expansionType": "alphabet", "depth": 0 },
{ "platform": "bing", "platformLabel": "Bing", "rank": 6, "relevanceScore": null, "suggestionType": null, "subtypes": null, "modifierQuery": "coffee maker a", "expansionType": "alphabet", "depth": 0 }
],
"scrapedAt": "2026-11-01T01:46:57.634Z",
"addedWords": ["alarm", "clock"],
"rankByPlatform": { "google": 3, "youtube": 4, "bing": 6 },
"entityName": null,
"entityDescription": null,
"ebayCategories": null,
"platformsText": "Google, YouTube, Bing",
"addedWordsText": "alarm clock"
}

Every keyword keeps its full provenance in sources - exactly which platform, rank, relevance score and modifier query surfaced it - so nothing the suggest APIs return is thrown away.

  • addedWords - the words the keyword adds on top of its seed ("coffee maker with grinder" → ["with", "grinder"]), ready for modifier clustering and negative-keyword lists.
  • rankByPlatform - the best autocomplete position per platform, flat and spreadsheet-friendly.
  • Entities - YouTube and Wikipedia suggestions that are real-world entities fill entity with name, description, entityId (a Knowledge Graph or Wikidata id), url, imageUrl and, for Wikipedia, pageId; the same name and description are repeated as the flat entityName / entityDescription columns.
  • ebayCategories - for keywords surfaced by eBay, the categories ({ "id": 20625, "name": "Kitchen, Dining & Bar Supplies" }) eBay associates with the query that surfaced them; each eBay entry in sources carries the same list as queryCategories.
  • platformsText / addedWordsText - the same platforms and added words as one line of text ("Google, Bing, Amazon", "with grinder"), so the table view and spreadsheets show the values directly.

Which fields are filled, and when

Every row carries the keyword, its seed, the platforms and ranks it surfaced on, the intent flags, the word metrics and the full sources list. Some fields come only from particular platforms, so they stay empty (null) on rows no such platform surfaced:

FieldFilled when
bestRelevanceScorethe keyword surfaced on Google (web, Images, News or Shopping): the only autocomplete that publishes a relevance figure. Amazon sends a placeholder 0, which is left empty rather than shown as a score.
entity, entityName, entityDescriptionthe keyword is a Wikipedia article title (most Wikipedia suggestions) or a YouTube suggestion that YouTube marks as a real-world entity (a few per search, such as a person or a song). Google's autocomplete no longer attaches entity details, so Google-only runs leave these empty.
ebayCategoriesthe keyword surfaced on eBay.
isQuestion, isCommercial and the other flagsalways filled (true / false), worked out from the keyword's own words.

Monitor mode - only new keywords

Turn on monitorMode and the actor remembers every keyword it has delivered (in a named key-value store keyed by monitorKey). On the next run it returns only keywords it has never delivered before - ideal for spotting emerging long-tail terms and trends without re-processing the same list. Keywords already delivered do not count toward maxResults, and a keyword left out by a cost limit or a time limit is delivered on a later run.

This is fully compatible with Apify Schedules: the schedule starts the run on your cron (e.g. daily), and monitor mode deduplicates against all prior runs. Use different monitorKey values to keep separate watch-lists per topic or market.

How much does it cost?

This Actor uses Apify's pay-per-event pricing: you pay for the keywords it delivers, with no monthly rental. The events it can charge are:

  • Keyword - one unique keyword with its seed, the platforms and rank where it surfaced, its search intent and word metrics, Google's relevance figure where Google gives one, and the full per-platform source list.
  • Actor start - from 17 November 2026, a small fee each time a run starts ($0.001 per GB of run memory, so $0.001 at the default 1 GB). It covers the cost of runs that end without a keyword, such as a filter that keeps nothing.

The current price of each event is shown on the Pricing tab of this page. Set a maximum total charge on the run if you want a hard cap on spend, and use the input limits to control how much the Actor fetches.

Automate & schedule

Run this actor on autopilot and pull results into your own stack:

import { ApifyClient } from 'apify-client';
const client = new ApifyClient({ token: 'MY_APIFY_TOKEN' });
const run = await client.actor('scrapesage/keyword-research-scraper').call({
keywords: ['running shoes'],
platforms: ['google', 'youtube', 'amazon'],
expansions: ['alphabet', 'questions'],
maxResults: 500,
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(`Got ${items.length} keywords`);

Integrate with any app

Connect the dataset to 5,000+ apps - no code required:

  • Make - multi-step automation scenarios.
  • Zapier - push new keywords straight into your content or ads workflow.
  • Slack - get notified when monitored keywords appear.
  • Google Drive / Sheets - auto-export every run to a spreadsheet.
  • Airbyte - pipe results into your data warehouse.
  • GitHub - trigger runs from commits or releases.

Use with AI assistants (MCP)

The output is clean, LLM-ready JSON. You can call this actor from Claude, ChatGPT, or any agent framework through the Apify MCP server - ask your assistant to "research long-tail keywords for coffee maker on Google, YouTube and Amazon" and let it run this scraper for you.

Agent-ready: autonomous payments (x402 & Skyfire)

This actor is agent-ready: AI agents can discover it, run it, and pay for it autonomously, with no Apify account and no human in the loop. It uses pay-per-event pricing and limited permissions, so it qualifies for Apify's agentic-payment standards:

  • x402: an open, HTTP-native payment protocol. Agents pay per run in USDC on the Base network directly through the Apify MCP server, with no account and no API key.
  • Skyfire: agent-to-service payments for fully autonomous AI-agent workflows.

Building an AI agent, MCP tool, or autonomous data pipeline? This scraper is ready to plug in and pay as it goes.

More scrapers from scrapesage

Build a complete content, SEO & e-commerce research stack:

Tips

  • Start small, then widen. Begin with a couple of platforms and ["alphabet","questions"]; add prepositions, comparisons, commercial and more platforms once you see the shape of the data.
  • Cross-platform = demand. Sort by platformCount to find keywords that show up on several platforms - those are your safest bets.
  • Filter for intent. Use searchIntent to split content keywords (informational) from money keywords (commercial / transactional).
  • Question content. Set includeQuestionsOnly to build FAQ and People-Also-Ask style pages.
  • Cost control. maxResults is a hard cap; recursiveDepth multiplies coverage fast, so raise it together with maxResults.

FAQ

Where does the data come from? From the public autocomplete / suggest endpoints each platform uses to power its own search box - the same suggestions you see when you start typing. No API key or login required.

Do you provide monthly search volume and CPC? No - those numbers are not exposed by autocomplete, and this actor does not fabricate them. Instead it gives you real signals: Google's relevance score, the suggestion rank, and how many platforms surface each keyword (platformCount) - an honest, cross-platform demand indicator.

Which platforms can I scrape at once? Any combination of the 14 supported platforms in a single run. With dedupeAcrossPlatforms on, the same keyword found on several platforms is merged into one row with a sources breakdown.

Can I research keywords in other countries and languages? Yes - set country and language as a code or a name. Google and YouTube use gl/hl, Bing and DuckDuckGo the matching region, TikTok its region and language, Yandex its language, Wikipedia the matching edition, and Amazon and eBay the store of that country (Amazon has 23 stores, eBay 14 country sites). Walmart is read in its US store and Naver for Korea. When the country has no store or market on one of them, that platform is skipped and the run's status message says so, rather than quietly reading another country's data.

Can I get TikTok keyword ideas? Yes - add tiktok to platforms. You get the suggestions TikTok's own search bar shows for your seed and its expansions, in the country and language you set, merged with the other platforms so you can see which TikTok searches also have demand on Google or YouTube.

How do I get only new keywords over time? Turn on monitorMode and run on a Schedule. Each run returns only keywords not seen in previous runs.

Can I export to Google Sheets, CSV, or Excel? Yes - one click in the dataset view, or automatically on every run via the Google Drive integration.

Is scraping autocomplete legal? This actor collects publicly available suggestion data only. You are responsible for using the data in compliance with applicable laws and each platform's terms.

Disclaimer

This Actor is an independent tool and is not affiliated with, endorsed by, or sponsored by Google, YouTube, TikTok, Amazon, Walmart and the other search engines and sites it queries, or any of their owners or subsidiaries. All trademarks mentioned are the property of their respective owners.

Site and brand names are referenced only in a descriptive, nominative sense - to identify the publicly accessible sources this Actor collects data from. It is not an official product of any of them, is not authorised or certified by them, and does not distribute their software. It collects only publicly available information; you are responsible for ensuring your use of that data complies with applicable laws, regulations and the terms of each source.

Need help?

Open an issue on the actor's Issues tab, or visit the Apify help center. Feature requests are welcome - this actor is actively maintained.