Google People Also Ask Scraper (Answers, 4 Levels)

Get every "People also ask" question Google shows for your keywords as a question tree up to 4 levels deep, each with its answer, source title and link, plus Google's AI Overview text with its cited sources, related searches and Autocomplete suggestions, as JSON or CSV rows. 44 countries, no login, no API key.

• Write content briefs: the real questions searchers ask about a topic, with the pages Google uses to answer them.
• Plan FAQ pages and headings: turn each question tree into sections that match what people search for.
• See if Google's AI Overview cites your site: add your domain to trackDomains and every overview row says whether it is cited.
• Find long-tail keywords: Autocomplete from A to Z and with question words gives hundreds of phrases per keyword.
• Track changes on a schedule: see when questions, overviews and citations appear, change or disappear.

Quick start

Paste this into the JSON tab of the input and click Start:

{"keywords": ["how to start a podcast"], "questionDepth": 1}

You get the questions Google shows for the keyword, the AI Overview with its sources, the related searches and about 15 Autocomplete suggestions: 28 rows in about 25 seconds, for about $0.03. Set "questionDepth": 2 to have each question searched too, which adds its answer and the questions it leads to.

More examples:

{"keywords": ["claim tax refund", "self assessment deadline", "marriage allowance"], "country": "gb", "questionDepth": 2, "autocompleteAlphabet": true}

Keyword ideas only, fast and cheap (no Google results pages are opened, so only the $1 rows apply):

{"keywords": ["crm", "email marketing"], "includePeopleAlsoAsk": false, "includeAiOverview": false, "includeRelatedSearches": false, "autocompleteAlphabet": true, "autocompleteQuestions": true}

What data can I get?

Every row has the same columns, so one table holds every result type.

Group	Fields
Every row	id (stable for the same keyword, type and text), keyword, type (paa, ai-overview, related, autocomplete), text, country, language, googleDomain, scrapedAt
Question tree	depth (1 = shown for your keyword, 2 = shown when a level-1 question was searched, and so on), parent (the search it appeared under), position
Answers	answer (the start of Google's answer), answerSource (ai-overview or top-result), answerTitle, answerUrl
AI Overview	text (Google's full overview as clean text, headings and list points kept as lines), sources (every cited page as {title, url, site})
Citation tracking	trackedDomainCited, citationState (cited, competitor-only, no-overview), citedTrackedUrls (only on runs with trackDomains)

• People Also Ask (paa): questions above the last level carry an answer. Google now writes the answer only when a question is clicked, so the Actor searches the question itself and takes the start of its AI Overview (with the first cited source), or the top result's snippet when there is no overview.
• AI Overview (ai-overview): one row per keyword searched. When Google shows no overview, text is empty and sources is [].
• Related searches (related): the "People also search for" phrases from every page searched.
• Autocomplete (autocomplete): suggestions for the keyword, and optionally for "keyword a" to "keyword z" and for "how keyword", "what keyword", "why keyword" and 10 more question words.

Questions are deduplicated within each keyword's tree, and a question that turns up under two keywords is searched only once.

One real AI Overview row and one question row (run on 2026-10-18, shortened):

[
  {
    "id": "f0828a52021d47a4",
    "keyword": "how to start a podcast",
    "type": "ai-overview",
    "text": "To start a podcast, you need to define your niche, pick your equipment, record and edit your episodes, and upload them to a hosting platform...\n1. Plan Your Concept and Niche\n- Choose a focused topic: ...",
    "depth": 0,
    "parent": "how to start a podcast",
    "position": 1,
    "sources": [
      {"title": null, "url": "https://creators.spotify.com/resources/how-to/start-a-podcast", "site": "creators.spotify.com"},
      {"title": "How to Start a Podcast: Complete Step-by-Step Guide [2026]", "url": "https://www.buzzsprout.com/blog/how-to-start-a-podcast", "site": "Buzzsprout"}
    ],
    "country": "us",
    "language": "en",
    "googleDomain": "www.google.com"
  },
  {
    "id": "d0c60db08396f247",
    "keyword": "how to start a podcast",
    "type": "paa",
    "text": "Do podcasts make money?",
    "depth": 1,
    "parent": "how to start a podcast",
    "position": 1,
    "answer": null,
    "country": "us",
    "language": "en"
  }
]

At questionDepth 1, level-1 questions have answer: null; at depth 2 or more they carry answer, answerSource, answerTitle and answerUrl. The dataset exports to JSON, CSV, Excel or HTML, and the run's SUMMARY record gives counts by type, citation totals and anything not finished.

How much does it cost?

You pay per row returned. There is no charge for Apify platform usage on top. In short: $1 per 1,000 rows, plus $4 per 1,000 keywords searched on Google and $4 per 1,000 answered questions.

Row	Free and Starter	Scale	Business
Result: a question without an answer, a related search or an Autocomplete suggestion	$1 per 1,000	$0.75 per 1,000	$0.55 per 1,000
Keyword searched on Google (its AI Overview row, empty when Google shows none)	$4 per 1,000	$4 per 1,000	$4 per 1,000
Answered question (each answer needs one more Google search)	$4 per 1,000	$4 per 1,000	$4 per 1,000

Plus $0.00005 per run start. Every Google results page costs real money to read, so each one is paid for by one $4 row. Apify's free $5 monthly credit covers about 5,000 results (questions, suggestions or related searches), or about 1,200 keywords searched on Google. On the Apify free plan, runs return up to 500 rows and make up to 25 Google searches each; any paid Apify plan removes both limits.

What's free: a keyword whose Google page comes back completely empty, rows skipped by onlyNew, and the Google search itself when Apify's Google SERP proxy can't be used on your account (the run says so). Autocomplete-only runs open no Google pages, so they cost only the $1 rows. To cap a run's spend, set Maximum cost per run in the run options, or maxResults in the input; the run stops cleanly at that point.

Limits, stated up front

• Google shows 4 or 5 questions per page. A deeper tree comes from searching each question: depth 2 typically adds 3 or 4 new questions per question. maxQuestionsPerKeyword (default 30) keeps it in check.
• Questions on the last level have no answer. Set the depth one higher to answer them.
• Not every keyword has an AI Overview. Many "how to", shopping and brand searches don't, and some countries show far fewer. Those keywords still get their row with empty text, because the search itself was made.
• Citation tracking reads the overview's cited links only: a brand named in the text without a link is not counted. Subdomains count (blog.example.com counts for example.com).
• Source links: Google hides destinations behind its own redirect links. The Actor asks Google where each one goes; if Google won't say, the row keeps Google's link, which still opens the page.
• Results are what a logged-out visitor in the chosen country sees, without personalisation.
• Speed: 10 Google pages are read at once and a page can take up to a minute. A keyword at depth 1 takes one page; depth 2 adds one page per question. A run that reaches its time limit stops cleanly with everything found so far and says which keywords were not finished.
• Up to 5,000 keywords per run.

Only new since last run, and schedules

Save a run as a task with "onlyNew": true and add a daily schedule. Each run returns every keyword's AI Overview row (a fresh read of Google, so you see when an overview appears, changes or disappears) and its answered questions, plus only the questions, related searches and suggestions earlier runs with the same country and language did not return. Skipped rows are free. Add trackDomains to see the day you gain or lose a citation. Set up a schedule under Schedules in Apify Console.

Integrations: API, n8n, Make and AI agents

API. Run it and get the rows back in one call:

curl -X POST "https://api.apify.com/v2/acts/locaihost~people-also-ask/run-sync-get-dataset-items?token=YOUR_APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"keywords": ["best crm for small business"], "trackDomains": ["example.com"], "includeAutocomplete": false}'

Apify's API clients for Python and JavaScript work the same way.

n8n, Make, Zapier, Google Sheets, Slack, webhooks. Apify's built-in integrations send each run's dataset wherever you need it, for example a content brief straight into a Google Sheet.

AI agents (MCP). Add this Actor to Claude, Cursor or any MCP client with https://mcp.apify.com/?tools=locaihost/people-also-ask, then ask something like "What questions do people ask Google about switching to a heat pump, and which sites does Google's overview cite?"

FAQ

How much does it cost to scrape People Also Ask?

One keyword at depth 1 with Autocomplete is about $0.03: one $4-per-1,000 keyword row plus about 27 questions, related searches and suggestions at $1 per 1,000. At depth 2, each question searched adds $0.004 plus the new questions it shows.

Can I use it through the API or from Python?

Yes. Use the run-sync-get-dataset-items call above, or Apify's Python and JavaScript clients. Every run's data is also available as JSON, CSV or Excel from its dataset.

Can I use it from an AI agent?

Yes. Through the Apify MCP server (https://mcp.apify.com/?tools=locaihost/people-also-ask) any MCP client can call it with a list of keywords and read the questions back.

Is it legal to scrape Google People Also Ask?

The Actor reads Google results pages that any logged-out visitor sees, through Apify's Google SERP proxy, and Google's public Autocomplete. It never logs in and collects no personal data. Google's terms restrict automated access, so check them against your use. You are responsible for how you use the data, including the copyright of the sites Google cites and the laws that apply to you. Do not use AI Overview text to train models. If you are a rights holder and want something changed, open an issue and we will act on it promptly.

Does my site appear in Google's AI Overview?

Add your domain (and rivals' domains) to trackDomains. Each AI Overview row then has trackedDomainCited, citationState and citedTrackedUrls, and the run's message and SUMMARY.citations give the totals. Nothing extra is searched, so the price is the same.

Why are some fields empty?

answer and its source fields are empty on the last level of the tree (raise questionDepth to fill them). text and sources on an AI Overview row are empty when Google shows no overview for that keyword. The citation columns exist only on runs with trackDomains, and are null on rows other than AI Overviews.

More from locaihost data

• Google Trends data: search interest over time and rising queries for the same keywords.
• SEO site audit: check your own pages' titles, headings and links before you write.
• YouTube transcripts: what creators say about a topic, as text.
• Website traffic checker: monthly visits and traffic sources for the sites Google cites.

Did it help?

If this saved you time, a quick rating on the Store helps other people find it.

Support

Something not working, a missing field or a keyword that reads wrong? Open an issue on the Issues tab: we answer issues within 24 hours.