Google AI Overview API

Google AI Overview API: the full answer and every source it cites

Expand the AI Overview from a Google search and get its full text in order, block by block, plus the list of sources it cites. €0.00075 per call, half the price of a search.

per 1,000 calls
€0.75
per call
€0.00075
0.075 cents
rows per call
up to 1
per 1,000 rows
≈ €0.75

Is Google's AI Overview quoting you or your competitor?

The Google AI Overview API loads the complete AI Overview for a Google search and returns it as one row. You get the answer split into ordered text blocks: paragraphs, headings and list items, with the links inside the text. You also get the cited sources with title, site name, link and the snippet Google took from each. One call costs €0.00075, so 1,000 overviews cost €0.75. That's half the price of a search call. It doesn't return the ranked results; those come from the search call. Every source and price is on our page of all data APIs and prices. Every overview you expand is saved in your Scrapewise account, so you can read it in the table, export Excel, ask your AI assistant or pull it by API.

It's a follow-up call. Google often loads the full AI Overview a moment after the results page, so a normal search only has part of it, or just a key to fetch the rest. This API takes that key and fetches the complete version. The key is short-lived, so run this call right after the search, in the same job.

Why track it? For many questions buyers ask, the AI Overview is the first thing on the page, above every blue link. If it cites three sites and yours isn't one of them, your position 2 ranking sits below an answer that doesn't mention you. Tracking which sources get cited, week by week, is the closest thing to rank tracking for AI answers.

Who uses the Google AI Overview API

  • SEO and content teams

    Check whether your pages are cited for your target keywords, and which competitor pages are cited instead.

  • Agencies reporting AI visibility

    Show clients their citation share across a keyword set before and after a content update, with dated rows as evidence.

  • Sellers and brands

    Read what the overview says about your product type or brand name. If it repeats an old price or a wrong claim, you see it in the text.

  • Researchers and writers

    Collect the sources Google trusts for a topic, then read those pages to see what they cover that yours doesn't.

How it works

  1. Run the Google search

    Call the Google search API for your keyword. When Google has an AI Overview to load, the results page carries a key for it.

  2. Expand the overview

    Pass the key to this API straight away. It returns the full overview text and sources for €0.00075, saved with its date in your database.

  3. Check the sources

    Search the source columns for your domain in the table or Excel, pull rows by API, or ask your AI assistant: 'which overviews this month cited my competitor and not me?'

Set it up once. The data keeps coming.

Save your search once and it runs by itself. Every run lands in your own Scrapewise database, and you pick how to read it.

  1. Set it up once

    Add your ASINs, keywords, places or apps to a scraper in the web app. New accounts get 5 free requests.

  2. Runs on your schedule

    Pick daily, weekly on the days you choose, every few days, or the first or last day of the month. Scheduled runs start at night, European time. Need it more often? Start runs from your own code.

  3. Saved in your database

    Each run adds dated rows, so this week sits next to last week. Rows are kept for 90 days.

  4. Use it your way

    Open the table in the web app and download the latest run as an Excel file. Ask your own AI assistant about it. Or pull the rows into your code with an API key.

Ask your AI assistant about your own data

Connect Claude Desktop, Claude Code or another MCP app with a read-only key. The assistant reads the rows you've already collected, at no extra cost on Scrapewise. With a full-access key it can start a run for you too.

  • Which of my ASINs lost the Buy Box this week?
  • Which competitor cut prices the most since Monday?
  • Show the keywords where I dropped out of the top 10.

What you send and what you get

Send the AI Overview key from a Google results page. You get one row per overview: its state, the text blocks in order and the cited sources.

You give
  • AI Overview session key (from a Google Search result)required
    ce5a60050323b4beee15b60e1fcc9018
You get

Up to 1 rows per call, 122 columns each

  • sw status
  • sw source url
  • sw title
  • sw url
  • sw seller
  • sw fetched at
  • state
  • text_blocks[0]_type
  • text_blocks[0]_snippet
  • text_blocks[0]_snippet_links[0]_text
  • text_blocks[0]_snippet_links[0]_link
  • text_blocks[1]_type
  • +110 more, see all columns

Sample rows from real AI Overviews

Real AI Overviews for US searches such as "how to start selling on amazon" and "what is the amazon buy box", collected on 16 September 2026. Each row is one overview. Scroll sideways to see all text blocks and sources.

5 sample rows, showing 15 of 122 columns.

5 sample rows, showing 15 of 122 columns.
sw status_sw_statusResult of this row: OK, no results, or an errorsw source url_sw_source_urlThe input you gave: a link, keyword or IDsw title_sw_titleMain name of the itemsw url_sw_urlDirect link to the itemsw seller_sw_sellerWho sells or publishes itsw fetched at_sw_fetched_atWhen we collected this row (UTC)statestateStatetext_blocks[0]_typetext_blocks[0]_typeText blocks #1 typetext_blocks[0]_snippettext_blocks[0]_snippetText blocks #1 snippettext_blocks[0]_snippet_links[0]_texttext_blocks[0]_snippet_links[0]_textText blocks #1 snippet links #1 texttext_blocks[0]_snippet_links[0]_linktext_blocks[0]_snippet_links[0]_linkText blocks #1 snippet links #1 linktext_blocks[1]_typetext_blocks[1]_typeText blocks #2 typetext_blocks[1]_snippettext_blocks[1]_snippetText blocks #2 snippettext_blocks[2]_typetext_blocks[2]_typeText blocks #3 typetext_blocks[2]_list[0]_snippettext_blocks[2]_list[0]_snippetText blocks #3 list #1 snippet
OKhttps://www.google.com/search?q=how+to+start+selling+on+amazon1. Choose a Selling PlanAmazon.com2026-09-16T09:55:51ZcompleteparagraphTo start selling on Amazon, visit Sell on Amazon to set up your seller account and choose a selling plan.Sell on Amazonheading1. Choose a Selling PlanlistIndividual Plan: Costs $0.99 per item sold. It works best if you plan to sell fewer than 40 items a month.
OKhttps://www.google.com/search?q=how+to+find+winning+products+to+sell+on+amazonResearching Marketplace DemandReddit2026-09-16T09:56:28ZcompleteparagraphAI OverviewheadingResearching Marketplace DemandlistCheck top lists: Visit the Amazon Best Sellers page to see what items sell fast in different categories.
OKhttps://www.google.com/search?q=what+is+the+amazon+buy+boxWhy It MattersHelium 102026-09-16T09:57:05ZcompleteparagraphThe Amazon Buy Box (officially renamed the Featured Offer by Amazon) is the prominent box on a product detail page where shoppers find the "Add to Cart" and "Buy Now" buttons.Featured OfferparagraphWhen a buyer clicks these buttons without choosing a specific alternative seller, the item from the seller currently winning the Buy Box is automatically added to their cart.heading
OKhttps://www.google.com/search?q=how+to+rank+a+youtube+videoSourcing ClipsYouTube2026-09-16T09:57:42ZcompleteparagraphAI OverviewheadingSourcing ClipslistFind trending HD clips with high engagement (over 10,000 likes) on TikTok or Instagram under a specific sub-niche (e.g., funny fishing moments, unique skills).
OKhttps://www.google.com/search?q=coffee+makerPopular Types and Prices2026-09-16T09:58:19ZcompleteparagraphA coffee maker is a kitchen appliance that automates brewing hot coffee by passing heated water through ground coffee beans.headingPopular Types and PriceslistDrip Coffee Makers: These classic machines brew large pots of coffee using a filter. Prices range from about $17 for a basic Mainstays Black 12-Cup Drip Coffee Maker Go to product viewer dialog for this item. up to $120 for the Cuisinart 14 Cup Programmable Coffee Maker Go to product viewer dialog for this item..

Column guide

One overview is one wide row. Text blocks are numbered in reading order from 0, and each block has a type: paragraph, heading or list. List items and inline links are numbered inside their block. Sources are numbered from 0 as well. The sw columns (shown as sw title, sw seller and so on) use the same names as on every Scrapewise data API.

Standard columns shared with every API

sw source url is the search the overview belongs to and sw title is the overview's main heading. sw seller is the first cited site, and sw fetched at is the time in UTC.

  • sw status
  • sw source url
  • sw title
  • sw url
  • sw seller
  • sw fetched at

Overview state

Shows whether the full overview loaded, for example complete.

  • state

Text blocks 1 to 4

The opening of the answer. Each block has a type and either a snippet (the text) or numbered list items. Inline links keep their anchor text and link.

  • text_blocks[0]_type
  • text_blocks[0]_snippet
  • text_blocks[0]_snippet_links[0]_text
  • text_blocks[0]_snippet_links[0]_link
  • text_blocks[1]_type
  • text_blocks[1]_snippet
  • text_blocks[2]_type
  • text_blocks[2]_snippet
  • text_blocks[2]_list[0]_snippet
  • text_blocks[2]_list[0]_snippet_links[0]_text
  • text_blocks[2]_list[0]_snippet_links[0]_link
  • text_blocks[2]_list[0]_snippet_links[1]_text
  • text_blocks[2]_list[0]_snippet_links[1]_link
  • text_blocks[2]_list[1]_snippet
  • text_blocks[2]_list[2]_snippet
  • text_blocks[2]_list[2]_snippet_links[0]_text
  • text_blocks[2]_list[2]_snippet_links[0]_link
  • text_blocks[2]_list[3]_snippet
  • text_blocks[2]_list[3]_snippet_links[0]_text
  • text_blocks[2]_list[3]_snippet_links[0]_link
  • text_blocks[3]_type
  • text_blocks[3]_snippet
  • text_blocks[3]_list[0]_snippet
  • text_blocks[3]_list[1]_snippet
  • text_blocks[3]_list[2]_snippet

Text blocks 5 to 10

The rest of the answer in the same shape. Longer overviews fill more of these columns; short ones leave them empty.

  • text_blocks[4]_type
  • text_blocks[4]_snippet
  • text_blocks[4]_snippet_links[0]_text
  • text_blocks[4]_snippet_links[0]_link
  • text_blocks[4]_snippet_links[1]_text
  • text_blocks[4]_snippet_links[1]_link
  • text_blocks[4]_snippet_links[2]_text
  • text_blocks[4]_snippet_links[2]_link
  • text_blocks[4]_list[0]_snippet
  • text_blocks[4]_list[1]_snippet
  • text_blocks[4]_list[2]_snippet
  • text_blocks[4]_list[3]_snippet
  • text_blocks[4]_list[4]_snippet
  • text_blocks[4]_list[5]_snippet
  • text_blocks[4]_list[6]_snippet
  • text_blocks[5]_type
  • text_blocks[5]_snippet
  • text_blocks[5]_list[0]_snippet
  • text_blocks[5]_list[1]_snippet
  • text_blocks[5]_list[2]_snippet
  • text_blocks[5]_list[3]_snippet
  • text_blocks[6]_type
  • text_blocks[6]_snippet
  • text_blocks[6]_list[0]_snippet
  • text_blocks[6]_list[1]_snippet
  • text_blocks[6]_list[1]_snippet_links[0]_text
  • text_blocks[6]_list[1]_snippet_links[0]_link
  • text_blocks[6]_list[2]_snippet
  • text_blocks[6]_list[3]_snippet
  • text_blocks[6]_list[4]_snippet
  • text_blocks[7]_type
  • text_blocks[7]_snippet
  • text_blocks[7]_snippet_links[0]_text
  • text_blocks[7]_snippet_links[0]_link
  • text_blocks[7]_list[0]_snippet
  • text_blocks[7]_list[1]_snippet
  • text_blocks[8]_type
  • text_blocks[8]_snippet
  • text_blocks[8]_list[0]_snippet
  • text_blocks[8]_list[1]_snippet
  • text_blocks[9]_type
  • text_blocks[9]_snippet
  • text_blocks[9]_list[0]_snippet
  • text_blocks[9]_list[1]_snippet
  • text_blocks[9]_list[1]_snippet_links[0]_text
  • text_blocks[9]_list[1]_snippet_links[0]_link

Cited sources

Up to 11 sources, each with the page title, link, the snippet Google took from it and the site name, such as Reddit or YouTube.

  • references[0]_title
  • references[0]_link
  • references[0]_snippet
  • references[0]_source
  • references[1]_title
  • references[1]_link
  • references[1]_snippet
  • references[1]_source
  • references[2]_title
  • references[2]_link
  • references[2]_snippet
  • references[2]_source
  • references[3]_title
  • references[3]_link
  • references[3]_snippet
  • references[3]_source
  • references[4]_title
  • references[4]_link
  • references[4]_snippet
  • references[4]_source
  • references[5]_title
  • references[5]_link
  • references[5]_snippet
  • references[5]_source
  • references[6]_title
  • references[6]_link
  • references[6]_snippet
  • references[6]_source
  • references[7]_title
  • references[7]_link
  • references[7]_snippet
  • references[7]_source
  • references[8]_title
  • references[8]_link
  • references[8]_snippet
  • references[8]_source
  • references[9]_title
  • references[9]_link
  • references[9]_snippet
  • references[9]_source
  • references[10]_title
  • references[10]_link
  • references[10]_snippet
  • references[10]_source

What teams do with AI Overview data

01

Track whether your site is cited

For 50 target keywords, check every week if your domain appears in the source columns. Chart the count over time.

02

See which competitors get cited

Count citations by site name across your keyword set. In our "what is the amazon buy box" sample, a seller tool blog and Amazon's own seller site are cited first.

03

Find content gaps

Read the headings and list items in the overview and tick off which points your page covers. The ones you miss are your next edits.

04

Report AI visibility to clients

Show citation share before and after a content update. Each row has the collection time, so the report is easy to back up.

05

Catch wrong claims about your brand

Save the overview text for your brand name and product names weekly and flag changes. An outdated fee or feature in the text is worth fixing at the source page.

Real setups and what they cost

  • Weekly citation tracking

    A content team tracks 50 keywords every week. About 217 overview calls cost €0.16 a month, plus €0.33 for the 217 Google searches that run first, so roughly €0.49 in total.

    calls per month
    217
    cost per month
    €0.163
  • Agency AI visibility report

    An agency checks 300 keywords for 4 clients once a month. The 1,200 overview calls cost €0.90, plus €1.80 for the 1,200 Google searches that run first.

    calls per month
    1,200
    cost per month
    €0.90
  • Daily brand watch

    A brand checks 10 questions about its products every day. 300 overview calls a month cost about €0.23, plus €0.45 for the searches.

    calls per month
    300
    cost per month
    €0.225

Why teams pick this API

  • The complete overview

    It fetches the version Google loads after the page, so you read the whole answer and the full source list.

  • Half the price of a search

    €0.00075 per overview. Tracking 50 keywords weekly, search included, costs under €0.50 a month.

  • Text you can search

    Blocks come in reading order with types, so finding your brand name or a wrong claim is a text search in a sheet.

What your AI Overview tracking would cost

Enter how many keywords you track and how often. Each overview is one call at €0.00075, on top of the €0.0015 search call that comes first. Compare all data APIs

Rows (up to)
1,000
You pay
€0.75
FAQ

Frequently asked questions

Google often loads the full AI Overview after the results page, so the search itself may only hold a key for it. This call uses that key to fetch the full overview. Keys expire quickly, so run it right after the search.

Ready to pull Google data into your stack?

Try it with 5 free requests. Then top up from €5 and pay €0.00075 per overview. See pay-as-you-go pricing for the details.