Long-Tail Keyword Finder: Google, YouTube, Amazon, Bing

One seed keyword in, hundreds of real long-tail phrases out. Read from the autocomplete of Google, YouTube, Bing, DuckDuckGo, Amazon and eBay at the same time, deduplicated into one table.

No API key, no account, no proxy.

Which of these keywords is actually growing

Autocomplete tells you what people type. It cannot tell you whether they are typing it more than they used to — so every seed is also checked against Google Trends, free, as part of the same run.

Each keyword row gains:

Field	Meaning
rising	Google Trends lists this keyword among the fastest-growing searches around the seed
risingGrowth	How fast, in Google's own words: +250%, or Breakout for more than +5,000%
isBreakout	The growth is above +5,000%
trendsRank	Its place in the most-searched list around the seed

And each seed summary gains trend (rising fast, rising, flat, falling, falling fast), trendChangePercent, risingKeywords and the breakout list.

Growth is measured against the same calendar weeks a year earlier, not against last quarter. That matters more than it sounds: read in September, padel looks like it is collapsing quarter on quarter, because it is a summer sport — against the same weeks last year it is growing.

It is measured against only the weeks that genuinely line up. Twelve months of weekly data reaches about two weeks past a full year, so that is a two-week comparison and comparisonPoints says so; reaching further back would put this September against last December, which is how a Christmas keyword ends up looking like a collapse in the autumn. Pick Past 5 years for a thirteen-week comparison. trendBasis always names which comparison was used.

Growing keywords that no autocomplete suggests are added to the results, labelled foundBy: "Google Trends". A recent run on "air fryer" added ninja glass air fryer (+700%), glass air fryer (+250%), mini air fryer (+200%) and best air fryer 2026 (Breakout) — none of which the autocompletes offered.

One honest caveat, because it is easy to get wrong: for some seeds Google's rising list drifts into generic shopping trends. "air fryer" came back with "laptop stand", "solar panels" and "hotel booking". Anything that does not carry a word from your seed is dropped before you see it.

What does this do?

Autocomplete is the only free, first-hand record of what people actually type. Every other keyword Actor reads one box — usually Google — and hands back a list. This one reads six, merges them, and tells you which platforms agreed.

flowchart LR
    A[Seed keyword] --> B[Expanded into ~450 queries<br/>a-z, questions, buying words]
    B --> C[Google · YouTube · Bing<br/>DuckDuckGo · Amazon · eBay]
    C --> D[One row per keyword:<br/>platforms, rank, intent]

Who is it for?

You are	You use it to
SEO and content writers	Build a topic cluster from the questions people really ask
YouTubers	See what the YouTube box suggests, which is rarely what Google suggests
Amazon and eBay sellers	Find the buying phrases shoppers type in the marketplace, per country
PPC managers	Harvest long-tail phrases and negative keyword candidates
App and product marketers	Find the words your market uses before you write the page
Agencies	Produce a keyword universe for a client brief in under a minute

What data do you get?

One row per keyword:

Field	Meaning
keyword, seed	The phrase, and the seed it came from
platforms, platformCount	Which boxes suggested it, and how many
bestRank	The highest position any platform gave it
timesSuggested	How many different queries surfaced it
score	Ranking signal: breadth, repetition and position combined
intent	question, comparison, local, commercial or general
wordCount, characterCount, isLongTail	Shape of the phrase (long tail is 4+ words)
containsSeed	Whether the seed survived, or autocomplete drifted to a neighbour

The run summary adds per-platform and per-intent totals, the top keywords and the top questions for each seed.

Example: three seeds, one run

A real run on 17 September 2026 from Apify, US market, 40 seconds, 4,089 keywords:

Seed	Keywords	Google	YouTube	Bing	DuckDuckGo	Amazon	eBay	Questions
running shoes	1,500	556	446	533	460	298	192	233
protein powder	1,500	587	528	596	484	354	160	231
project management software	1,089	445	313	488	327	7	0	117

That last row is the point of reading six boxes: a software term has no shopping demand at all, while "running shoes" has 490 marketplace phrases that Google never shows you.

What this does not give you, and why

There is no search volume and no cost per click. Those numbers come from Google Ads, which needs a paid account and an API key, and this Actor is built to run with neither. Anything else you see sold as "free search volume" is a scrape of somebody's dashboard and breaks when they change it.

Instead you get three honest popularity signals straight from the source:

• platformCount — six boxes suggesting the same phrase is a strong signal
• bestRank — autocomplete lists are ordered by popularity, so position 1 means something
• timesSuggested — how many different queries pulled it back

Also worth knowing: the country behind Google, YouTube, DuckDuckGo and Bing follows the server, which is in the United States. You can pick the language, and the shopping market setting really does switch Amazon, eBay and Bing to another country, but "Google suggestions as seen from Germany" is not something this can promise.

How to use it

1. Put your seed keywords in, one per line. Short seeds return more.
2. Choose the platforms. Drop Amazon and eBay for services and software; keep them for anything physical.
3. Choose the expansions: letters a-z, question words, buying words, prepositions.
4. Set the language and, for the shopping platforms, the market.
5. Run it. One seed with all six platforms is about 450 requests and takes 5 to 25 seconds.

Pay-per-event pricing

Event	Price
Seed keyword	$0.01
Keyword returned	$0.0002

One seed returning 1,000 keywords costs $0.21. Three seeds and 4,089 keywords: $0.85.

Popular keyword sets

• YouTube Keyword Research: What Viewers Search For — "how to edit videos" on youtube
• Amazon Keyword Research: What Shoppers Actually Type — "protein powder" on amazon
• eBay Keyword Research for Listings and Titles — "vintage watch" on ebay
• Google Autocomplete Keywords, Expanded A to Z — "running shoes" on google
• Question Keywords for Blog Posts and FAQ Pages — "coffee maker" on google, youtube, bing, duckduckgo
• Long-Tail Keywords for SEO From Four Search Engines — "project management software" on google, youtube, bing, duckduckgo
• Fitness Keyword Research: Gear, Workouts, Questions — "home gym equipment" on all six platforms
• Skincare Keyword Research for Content and Listings — "face serum" on all six platforms
• Coffee Keyword Research: Gear, Brewing and Buying — "espresso machine" on all six platforms
• Pet Supplies Keyword Research for Stores and Blogs — "dog food" on all six platforms
• Travel Keyword Research for Blogs and Gear Stores — "carry on luggage" on all six platforms
• Gaming Keyword Research for Channels and Stores — "gaming chair" on all six platforms
• Software Keyword Research for SaaS Content and Ads — "crm software" on google, youtube, bing, duckduckgo
• Home and Office Keyword Research for Stores — "standing desk" on all six platforms
• UK Keyword Research: Amazon, eBay and Bing UK — "running shoes" on amazon, ebay, bing

Each page is the same Actor with a different seed and platform set. Open one, put your own keyword in, and run it.