Keyword Trends Finder – Autocomplete Ideas + Google Trends

Turn one seed keyword into hundreds of keyword ideas, and see at a glance which ones are rising. The Actor combines Google Autocomplete (seed + a–z, question words and modifiers) with Google Trends (rising and breakout related queries) and checks the trend direction of each idea: rising, stable or falling, with the change in %.

$3 per 1,000 keyword ideas. One flat, CSV-ready row per keyword, ready for your SEO tool, spreadsheet or content calendar.

Why this Actor

Most keyword tools give you a long list. This one tells you where demand is going. Example from a test run (ai agents, US, 12 months):

keyword	source	trend
ai agents	seed	rising, +23.5 %
openclaw	trends-rising	Breakout (kept when also found by Autocomplete or listed in relevanceKeywords)
consumer adoption of ai agents	trends-rising	+1,150 %
ai agents explained	autocomplete	falling, −50.5 %

What you get for each keyword

Field	Meaning
keyword, seedKeyword	the idea and the seed it came from
source	seed, trends-rising (rising/breakout on Google Trends), autocomplete, trends-top
modifier, suggestionRank	which expansion found it (e.g. how, vs, b) and its position in Google's suggestions
trendDirection	rising, stable, falling, no-data, or not-checked (beyond the trend-check limit)
trendChangePercent, trendAverage, trendLatest, trendPeakDate	trend details for the selected period
trendsValue, isBreakout	growth shown by Google Trends for rising queries ("+250 %", "Breakout")
isQuestion, wordCount	quick filters for FAQ content and long-tail keywords
trendsUrl, googleSearchUrl	open the keyword on Google Trends or Google Search

Use cases

• SEO and content planning: find long-tail and question keywords, and write first about the ones that are rising.
• Product and niche research: see which sub-niches grow and which fade.
• Ads: spot breakout searches before the cost per click goes up.
• Local markets: any country and language. German uses German question words (wie, was, warum …).

How to use

1. Add one or more seed keywords.
2. Choose country and language, and how many ideas you want per seed (up to 500).
3. Keep Add Google Trends direction on, and choose how many ideas get a trend check (default 30 per seed).
4. Run it, then export as CSV, Excel or JSON, or connect to Google Sheets, Make, Zapier or n8n.

{
  "seedKeywords": ["ai agents"],
  "country": "us",
  "language": "en",
  "maxKeywordsPerSeed": 100,
  "maxTrendsPerSeed": 30
}

Relevance of rising and top queries

filterRelevantTrends is on by default. Before applying the keyword limit, checking candidate trends or charging for output, the Actor keeps a Google Trends related query if:

• it contains all meaningful word stems from its seed (whole words, case-insensitive, any order), or
• the same query was independently returned by Google Autocomplete for that seed's expansions, or
• it contains all meaningful word stems of one entry in your optional relevanceKeywords list, or
• it matches a related term from a configured topic activated by that seed, such as ollama for local llm.

Common English and German question words are ignored. Simple English s-plurals and a small explicit list of English/German word forms are normalized: ai agent keeps ai agents, and ki agent keeps ki agenten. It is not a general stemmer or semantic model. Irregular plurals, compounds, other languages and ambiguous names can still be missed or matched incorrectly. Word boundaries remain required (ai does not match air). A new product name such as openclaw can still be discovered through Autocomplete. To keep a known synonym or product without that evidence, add it to relevanceKeywords. These extra user anchors apply to every seed; use separate runs for different topics. Built-in topic aliases apply only when that seed matches the topic's anchor. No extra network requests are made.

Maintainers configure word forms, plural exceptions and topic aliases in the Actor-owned src/relevanz_standard.json. This file also contains fallback stopwords. The shared src/einstellungen.json may override relevanz.stopwoerter; if that key or section disappears during a shared rollout, the Actor uses its own standard stopwords and writes a warning to the run log. Include the Actor-owned JSON file when deploying this change.

For example, the filter removes cheapest flights to tokyo, how to bake a cake and best time to visit maldives for ai agents, local llm and self hosted ai, unless explicitly allowed or independently returned by Autocomplete. Breakout status or a high growth percentage alone is not evidence of relevance. The run log reports how many unique Trends queries per seed were filtered.

Seed and Autocomplete rows are unchanged. With incomplete Autocomplete data, synonyms and unfamiliar names may be omitted from Trends results. Set filterRelevantTrends to false for broad discovery with the original ordering, limits and output fields. Fewer results may be returned when unrelated queries are removed. Only returned keyword rows are charged.

{
  "seedKeywords": ["ai agents"],
  "filterRelevantTrends": true,
  "relevanceKeywords": ["openclaw"]
}

Pricing

Pay per event: $0.003 per keyword idea returned ($3 per 1,000). Trend checks, platform usage and proxy are included.

Use with AI agents (MCP)

This Actor works as a tool for AI assistants and agents – Claude, ChatGPT, Cursor, VS Code, n8n and other MCP clients – through Apify's hosted MCP server. Add this server URL to your client:

https://mcp.apify.com?tools=stevenkramp/keyword-trends-finder

Sign in with your Apify account when asked. Your agent can then call the Actor in plain language, for example: "Find 100 keyword ideas for 'air fryer recipes' in the UK and show me only the rising ones." Runs started by your agent are normal Actor runs on your Apify account at the same pay-per-event price.

More from stevenkramp

Other Actors by the same developer – same quality standards, pay only for results:

Search & trends

• Google Trends Scraper – interest over time, regions, rising queries
• Google News Scraper – news articles with real URLs
• Google Images Scraper – full-size image URLs
• Google Shopping Scraper – prices and merchants
• Google Jobs Scraper – job listings

Apps

• Google Play Store Scraper – Android app data and rankings
• Google Play Reviews Scraper – Play Store reviews and ratings
• Apple App Store Scraper – iPhone, iPad and Mac app data
• Shopify App Store Scraper – Shopify apps and pricing plans

Research & media

• arXiv Papers Scraper – research papers and abstracts
• Apple Podcasts Scraper – podcasts with latest episodes

Websites & places

• Website SEO Audit – broken links, titles, sitemap
• Germany Neighborhood Profile – German neighborhood rents and vacancy

FAQ

Why do some keywords say not-checked? Trend checks take a few seconds each. Raise "Trend checks per seed" (up to 200) to check more.

What does no-data mean? Google Trends has too little search volume for that keyword. Very specific long-tail keywords often have no trend data.

Need full Google Trends data (regions, timelines, comparisons)? Use our Google Trends Scraper.

Personal data? None. Only public, anonymous search suggestions and Google Trends statistics.

Something broken? Open an issue in the Issues tab. Google changes its pages from time to time, and we fix scrapers quickly.