Google Autocomplete Scraper

Why use this Actor?

Google Autocomplete keyword scraper for long-tail SEO research. Expand seeds with A–Z, numbers, questions, prepositions, comparisons and custom modifiers, recurse through suggestions, run country/language matrices, cap requests/results. From $0.50 per 1,000 keywords, platform usage included.

Features

• Seed keywords — Seed queries to expand with Google Autocomplete.
• Language — Google hl language code, e.g. en, it, de.
• Country — Google gl country code, e.g. US, IT, GB.
• Suggestion sources — Which Google autocomplete to read. Web is the classic Google search box; YouTube returns video-search suggestions; Shopping returns product-search suggestions (often fewer). Each extra source repeats every query, so requests and results scale with the number of sources. A suggestion found in two sources is a separate result in each.
• Suggestion endpoint — Extended returns up to 50% more suggestions for the same number of requests, plus Google's relevance score (relevance) and suggestion type (suggestType, for example QUERY or ENTITY). Website-link suggestions it returns are skipped and never billed. Standard keeps the exact behaviour of earlier runs.
• Append A-Z — Query seed + each letter a-z.
• Prepend A-Z — Query each letter + seed for extra discovery.
• Append 0-9 — Also query seed + digits 0-9 for numeric long-tail variations.
• Question prefixes — Add question prefixes: how/what/why/when/where/who/which/can/does/is/best (or their translations when "Localized modifiers" is on).
• Maximum unique suggestions — Hard cap on unique paid suggestion rows emitted across all seeds.
• Maximum Google requests — Hard cap on Google Suggest requests for predictable runtime and request volume.
• Delay between requests (ms) — Delay between Google Suggest requests to reduce burst rate.

Use cases

• Seo keyword research.
• Long-tail discovery.
• Content planning.
• Search-intent and trend monitoring.

Related tools

• Google Trends Scraper - Keywords, Regions & Trending
• Google Ads Transparency Scraper
• Google News Scraper API

Example input

{
  "seeds": [
    "artificial intelligence"
  ],
  "language": "en",
  "country": "US",
  "sources": [
    "web"
  ],
  "suggestClient": "firefox",
  "expandAlphabet": true
}

Pricing & cost control

The primary event costs $0.000500 per Unique autocomplete suggestion (about $0.50 per 1,000 successful primary events). Only successful primary events are intentionally billed by this Actor; summary/status rows add context without adding primary-event charges.

Use the bounded input limits and filters to keep both event charges and platform usage predictable.

FAQ

What is this Actor for?
It is designed for SEO keyword research, long-tail discovery, content planning.

Can I run it on a schedule?
Yes. You can schedule Actor runs on Apify and send the resulting dataset into automations, webhooks, storage, or downstream APIs.

How do I control cost and run size?
Use the input limits and filters shown in the Actor input form. The Actor applies bounded defaults and hard caps so large jobs remain predictable.

Pricing at a glance

	Price per 1,000 unique keyword suggestions
This Actor	$0.50
Median of 8 comparable Apify Store Actors (October 2026)	$1.00

That is about 50% below the median per-result price for this task in Apify Store.

Subscriber discounts: on a paid Apify plan you pay even less per result — Bronze −10%, Silver −15%, Gold and higher −20% ($0.40 per 1,000 unique keyword suggestions on Gold).

Apify platform usage (compute, storage, data transfer) is included — you pay only the result price plus a small run start fee ($0.0001 per GB of run memory). No surprise compute bill, even on long runs.

Use the run's maximum charge setting to cap spending — the Actor stops cleanly when the limit is reached.

Use it from AI agents (MCP) and code

AI agents — Claude, ChatGPT, Cursor, VS Code, n8n AI agents and any MCP client can run this Actor through the official Apify MCP server: add https://mcp.apify.com?tools=zenomastro/google-autocomplete-keywords to your client. Then just ask, for example:

"Give me long-tail Google autocomplete keyword ideas for 'web scraping' in the US, including questions."

The agent fills the input from the field descriptions and receives clean JSON it can reason over.

API — one HTTP call runs the Actor and returns the results directly:

curl -X POST "https://api.apify.com/v2/acts/zenomastro~google-autocomplete-keywords/run-sync-get-dataset-items?token=YOUR_APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"seeds": ["apify"], "expandAlphabet": false, "expandQuestions": false, "expandPrepositions": false, "maxSuggestions": 5, "maxSuggestionsPerSeed": 5, "maxRequests": 2, "delayMs": 0}'

The same works from the Apify Python/JavaScript clients, Make, n8n, Zapier, LangChain and LlamaIndex integrations, or on a schedule.

What does Google Autocomplete Scraper do?

It turns a few seed keywords into hundreds of long-tail keyword ideas straight from Google's own search-box suggestions. Each seed is expanded with A–Z letters, numbers, question words, prepositions, comparisons and your own modifiers, optionally recursively, in any language and country, for Google web search, YouTube or Google Shopping. Every suggestion comes back as a clean row with its rank, search intent, keyword tags and word count, ready for SEO, PPC, content planning or AI workflows. It uses Google's public Suggest endpoint directly: no browser, no API key, fast and cheap.

Who it is for: SEO specialists and content marketers building keyword lists and FAQ sections, PPC managers looking for new search terms, e-commerce and YouTube creators researching what people type, agencies running local keyword research in several countries, and developers or AI agents that need keyword ideas as JSON.

How to use it

1. Enter one or more seed keywords, e.g. running shoes.
2. Pick the language and country (or a language × country matrix) and the suggestion source: Google web search, YouTube or Google Shopping.
3. Choose the expansion patterns: A–Z (on by default), questions and prepositions (on by default), numbers, comparisons, prefix letters, your own prefixes and suffixes, and optionally a recursive depth of 2–3.
4. Set the limits and, if you like, the free filters: exclude words, keep only long-tail suggestions or only certain tags (e.g. only questions). Filtered suggestions are never charged.
5. Click Start and export the results as JSON, CSV, Excel or HTML, or read them through the API. Save the input as a task and schedule it with a monitor name to track new and disappeared suggestions.

Main features

• Every common expansion pattern: seed + a–z, a–z + seed, seed + 0–9, question prefixes, prepositions, comparisons/alternatives, custom prefixes and suffixes, and recursion depth 1–3 (suggestions become new queries).
• Localized patterns: with localizeModifiers, question, preposition and comparison words are used in Italian, German, Spanish, French, Portuguese or Dutch.
• Many markets in one run: up to 20 languages × 20 countries (100 combinations), and web, YouTube and Shopping suggestions.
• Keyword intelligence on every row, free: searchIntent (transactional, commercial, navigational, informational), tags (question, comparison, local, price, free, best, review, how_to, year, long_tail), isQuestion, questionWord, wordCount, Google's sourceRank and, with the extended endpoint, Google's relevance score.
• Pay only for what you want: exclude words (excludeTerms), require a minimum number of words (minWordCount), keep only chosen tags (onlyTags), and deliver a suggestion shared by several seeds only once (dedupeAcrossSeeds). Duplicates and filtered suggestions are skipped before saving and never charged.
• Free seed insights: one seed_summary row per seed, language, country and source with the number of unique suggestions, the most frequent accompanying words (topTerms, the seed's themes), question count, tag and intent counts.
• Keyword monitor: with monitorName, every row gets isNew, previousRank and rankDelta, the seed summary lists added and removed suggestions, and onlyNew saves and charges only new suggestions.

Input examples

Long-tail research for one product, questions included:

{
  "seeds": ["running shoes"],
  "language": "en",
  "country": "US",
  "expandAlphabet": true,
  "expandQuestions": true,
  "expandPrepositions": true,
  "expandComparisons": true,
  "maxSuggestions": 1000
}

Only questions for an FAQ page, in Italian, without job-related ideas:

{
  "seeds": ["scarpe da corsa"],
  "language": "it",
  "country": "IT",
  "localizeModifiers": true,
  "onlyTags": ["question"],
  "excludeTerms": ["lavoro", "offerte di lavoro"],
  "maxSuggestions": 300
}

Several overlapping seeds in two markets, each idea delivered once:

{
  "seeds": ["crm software", "crm for small business", "best crm"],
  "languages": ["en"],
  "countries": ["US", "GB"],
  "dedupeAcrossSeeds": true,
  "minWordCount": 3,
  "recursiveDepth": 2,
  "maxRequests": 1500
}

YouTube keyword ideas with the extended endpoint (more suggestions per request plus Google's relevance score):

{
  "seeds": ["home workout"],
  "sources": ["youtube"],
  "suggestClient": "chrome",
  "maxSuggestions": 500
}

Output

Each keyword idea is one row (recordType: "suggestion"). Example with illustrative values:

{
  "recordType": "suggestion",
  "success": true,
  "sourceIndex": 0,
  "seed": "running shoes",
  "query": "how running shoes",
  "queryDepth": 1,
  "expansionMode": "question",
  "modifier": "how",
  "suggestion": "how to choose running shoes",
  "sourceRank": 1,
  "suggestSubtypes": [512],
  "searchIntent": "informational",
  "wordCount": 5,
  "isNew": null,
  "previousRank": null,
  "rankDelta": null,
  "language": "en",
  "country": "US",
  "source": "web",
  "relevance": null,
  "suggestType": null,
  "modifierLanguage": "en",
  "fetchedAt": "2026-10-14T07:00:01.204Z",
  "tags": ["question", "how_to", "long_tail"],
  "isQuestion": true,
  "questionWord": "how"
}

Field	Meaning
seed, query, expansionMode, modifier, queryDepth	Which seed and which exact query (pattern and depth) produced the suggestion
suggestion, sourceRank	The keyword idea and its position in Google's list (1 = first)
searchIntent	transactional, commercial, navigational, informational or mixed_or_unknown
tags, isQuestion, questionWord	Local keyword tags; see the list below
wordCount	Number of words
language, country, source	Locale and suggestion source (web, youtube, shopping)
relevance, suggestType	Google's relevance score and suggestion type (extended endpoint only)
isNew, previousRank, rankDelta	With a keyword monitor

Tags are computed locally from the words of the suggestion, in English plus the row's language when it is Italian, German, Spanish, French, Portuguese or Dutch: question (starts with a question word or ends with "?"), comparison (vs, alternative, competitors…), local (near me, nearby, open now…), price (price, cost, cheap, discount…), free, best (best, top 10…), review (review, rating, legit…), how_to (how to, guide, tutorial…), year (contains a year such as 2026) and long_tail (4 or more words).

Free rows that are never charged:

• seed_summary: one per seed, language, country and source with requestsUsed, failedRequests, uniqueObserved, rowsEmitted, complete, topTerms, questionCount, tagCounts, intentCounts, filteredOut, duplicatesSkipped, verbatimRelevance, urlSuggestionsSkipped and, with a monitor, added/removed lists and counts. The insights count every unique suggestion Google returned for the seed, also ones not delivered because of caps, filters or onlyNew.
• summary: totals, stoppedReason, filteredOut, crossSeedDuplicatesSkipped, throttleEvents, warnings and monitorStateSaved.
• error: a failed request (with its query, language, country and source) or an input problem.

Reliability

• A failed Google request becomes a free error row and the run continues; a run never fails because of one bad query or a typo in the input (input problems end with one clear error row).
• Temporary errors are retried with backoff. When Google answers HTTP 429 or 503, the Actor honours Retry-After, waits longer and slows its pace (up to 2 seconds between requests), then speeds up again after 20 clean requests. HTTP 400/404 answers are not retried.
• If Google rejects 10 requests in a row, the run stops early and says why (stoppedReason: "consecutiveFailures").
• Shortly before the run timeout the Actor stops and saves its summaries and monitor state (stoppedReason: "timeLimit").
• At your maximum charge limit the run stops cleanly; suggestions that did not fit are not marked as seen, so a monitor delivers them next time.

Limits

• Google returns up to 10 suggestions per request (up to 15 with the extended endpoint).
• Up to 100 seeds, 100 language × country combinations, 5,000 requests and 100,000 suggestions per run.
• Tags and search intent are rule-based keyword matching, not a language model; they are fast and free but can miss unusual phrasings.
• Google's Suggest endpoint is public but undocumented and may change its behaviour or rate limits; keep the default delay for large jobs.

FAQ

How many keywords will I get?
It depends on the seed and the patterns. With the default patterns (A–Z, questions, prepositions) one seed sends 47 requests, so up to about 470 suggestions before duplicates are removed. Recursion depth 2 or 3 sends one more request per suggestion found, so the request cap usually decides the run size. maxSuggestions, maxSuggestionsPerSeed and maxRequests cap every run.

Do I pay for duplicates or filtered suggestions?
No. Only saved suggestion rows are charged. Duplicates within a seed, suggestions removed by excludeTerms, minWordCount or onlyTags, suggestions skipped by dedupeAcrossSeeds or onlyNew, website-link suggestions and all summary and error rows are free.

Does it give search volume or CPC?
No. Google Autocomplete does not publish volumes. Use sourceRank and relevance as popularity signals, or combine the keywords with Google Trends data.

Can I get only questions for an FAQ or People-Also-Ask style list?
Yes: keep expandQuestions on and set onlyTags to ["question"].

How do I track keyword changes over time?
Save the input as a task with a monitorName, schedule it, and read isNew, rankDelta and the seed summary's added and removed lists.

Something is missing or not working?
Open an issue on the Issues tab with the run ID; feature requests are welcome.

Extended capabilities

• Run a bounded language × country matrix in one job.
• Combine A–Z, numeric, question, preposition, comparison, recursion, and custom prefix/suffix expansions.
• Track new/removed suggestions plus previous/current source rank and rank deltas.