What does the Google Search Results tool do?
Google Search Results runs a live Google search for your query and returns what the page shows, in order. You get each organic result with its position, title, URL, site name, snippet and date, and the parts of the page around them: the AI Overview and the pages it cites, People also ask, related searches, popular products and expert perspectives. Choose the country, language, device and location, and the result reflects that searcher instead of your own browser.
You ask
Show the Google results for “best trail running shoes” in the US, with position, site and date for each organic result.
It runs
From your AI assistant over MCP or from your code over REST
You get
- Organic results
- AI Overview
- Questions and searches
- Products and perspectives
- Search details
- Plan content from the real pageSee which formats rank for a query before you choose between a guide, a comparison and a product page.
- Check positionsFind where a domain appears for a query on a given day, in a given country and on a given device.
- Watch AI OverviewsSee whether a query triggers one and which pages it cites.
- Collect real questionsPull People also ask and related searches for FAQ sections and new topics.
- Give research agents a search stepLet Claude or ChatGPT search Google, pick sources and read them in the same chat.
What data comes back from a Google search?
One run returns the whole first page, split into parts. These are the fields our test run returned, described in plain words.
Organic results
- Position on the page
- Title
- Page URL
- Displayed link and site name
- Snippet with highlighted words
- Date, when Google shows one
- Star rating from rich results
- Length and preview for video results
AI Overview
- Summary text and lists
- Cited sources with title, link and snippet
- Which sources each statement relies on
Questions and searches
- People also ask questions
- Related searches
- AI Mode follow-up questions
Products and perspectives
- Popular products with price, rating and review count
- Discounts and original prices
- Perspective articles with source and date
Search details
- Query as Google displayed it
- Google results URL
- Time taken
- Next-page token and links to later pages
Parts of the page appear only when Google shows them, so a query without an AI Overview returns none. The result lists pages, not their text: open any URL with Page Reader to read it in full.
Sample output from a real search
The table at the top of this page is a real run from 16 November 2026: “best trail running shoes”, United States, English, desktop. It came back in under 9 seconds with 8 organic results, an AI Overview, 4 People also ask questions, 8 related searches, 8 popular products and 3 perspectives. Below the chart is everything on that page beyond the organic list.
One result in full
The rest of the same results page
- AI Overview
- Shown, citing 8 sources: 4 web pages and 4 product listings
- Pages cited in the overview
- reddit.com, runrepeat.com, trailandkale.com, brooksrunning.com
- People also ask
- 4 questions, such as “Which brand is best for trail running shoes?”
- Related searches
- 8, such as “Best trail running shoes for wide feet”
- Popular products
- 8, priced $112.50 to $169.99, with rating and review count
- Perspectives
- 3 articles, from runrepeat.com, irunfar.com and bettertrail.com
- AI Mode follow-ups
- 6 suggested questions
- More results
- Token for the next page, links to pages 2 to 9
What this sample tells you
The 8 organic results split into 3 editorial reviews, 2 YouTube videos, 2 community pages (a forum thread and a Facebook answers page) and 1 retailer category. Most of the page compares shoes, and only one result sells them.
5of 8
5 of 8 results show a date, and 3 of those 5 are four or more years old: a 2017 forum thread, a 2021 newspaper review and a video posted 4 years ago. Older pages still hold top spots here.
The AI Overview cites 4 web pages and 4 product listings. None of those 4 web pages is among the 8 organic results, so track the overview's sources separately from rankings.
Popular products on the same page ranged from $112.50 to $169.99, a quick read on the price band shoppers compare for this query.
How to search Google with your AI assistant
Connect AI/ML API
Click Try it now, pick your client and add the MCP server. You sign in with your AI/ML API account once.
Give the query and the market
Type the query as a searcher would, plus the country, language and device you care about.
Read the page
Your assistant returns the organic results in order, with the AI Overview, questions and related searches.
Act on it
Ask it to label page types, find a domain's position, or open the top three results and compare them.
Prompts to start with
Use the AI/ML API MCP Google Search Results tool. Search Google in the US for “best trail running shoes” and show the organic results with position, title, site and date, plus the sources cited in the AI Overview.
Use the AI/ML API MCP Google Search Results tool. Check whether rei.com appears in the first 20 Google results in the US for “waterproof hiking boots” on mobile, and at which position.
Use the AI/ML API MCP Google Search Results tool. Collect the People also ask questions and related searches for “standing desk benefits” and group them into FAQ topics.
Use the AI/ML API MCP Google Search Results tool. Search Google News for “standing desk” from the past month and list each headline with its site and date.
Search settings that change your results
Google tailors results to the searcher, so the settings matter as much as the query. In chat your assistant fills these in from your request. In code you pass them yourself.
| Setting | What it controls | Tip |
|---|---|---|
| Query | The search words, up to 2,048 characters | Write it the way a person types it into Google |
| Country | Two-letter country code for the results, such as us or gb | Match the market you sell in |
| Language | Interface and result language, such as en or de | Keep it fixed when you compare runs over time |
| Location | A city or region such as “Austin, Texas”, or latitude and longitude with a radius | Set it for any query with local intent |
| Device | Desktop, tablet or mobile layout | Check mobile separately when most of your traffic is mobile |
| Results and pages | Up to 10 results per page, then a token for the next page | Ask for page 2 when you track positions beyond the top 10 |
| Date filter | Past day, week, month or year, with a count such as 3 months | Use it to see only recent coverage of a topic |
| Site filter | Limit results to one site, or exclude one | See which of your own pages Google shows for a topic |
| Vertical | Local, video, news, shopping or patents instead of web results | Use news for coverage checks and video for YouTube research |
| Organic only | Skip the AI Overview and other features and return organic results only | Use it when you only need the list of ranking pages |
| Personalization and autocorrect | Ask for non-personalized results and turn off automatic spelling fixes | Turn both off for repeatable position checks |
Google mixes formats on one page. In our run the 8 organic results sat alongside an AI Overview, 8 popular products and 3 expert perspectives, so read the whole page before you decide what to write.
Ways teams use Google search data
Content briefs
See which formats rank, such as reviews, videos, forums or shops, and plan the page type that fits the query.
TitleURLSnippet
Position checks
Find where your domain appears for a list of queries on a given day, with country, language and device pinned.
PositionURL
AI Overview monitoring
Check whether your core queries show an AI Overview and whether your pages are among its sources.
AI OverviewCited sources
Question research
Collect People also ask questions and related searches to fill FAQ sections and find new topics.
People also askRelated searches
Competitor watch
Run your main queries every week and note when a new site reaches the first page.
PositionSite nameDate
Sources for an agent
Let your assistant search, pick the strongest results and read them in full before it writes anything.
URLSnippet
Google Search Results in your code
In your code
REST · your AI/ML API key, JSON records back
The same tool runs over REST with your AI/ML API key. Send a query with a country, language and device and you get JSON with the organic results in order, plus separate sections for the AI Overview, People also ask, related searches, products and perspectives. A token in the response fetches the next page.
Use MCP when a person researches in a chat and REST when a scheduled job checks positions or collects questions for a dashboard. Both use one AI/ML API account and one balance, and you skip running your own browsers and proxies. Start with the Toolbox REST quickstart.
Other ways to get Google results
Each route answers a slightly different question. The right one depends on whether you need the page a real searcher sees, a history of positions or a search box for your own sites.
| Google's programmable search API | Rank tracker in a paid SEO suite | Your own scraper | ||
|---|---|---|---|---|
| Setup | Connect once with your AI/ML API account | Google Cloud project, API key and a search engine ID | Subscription, project setup and a fixed keyword list | Browser automation, proxies and parsing to maintain |
| Inside an AI chat | Yes. Over MCP | Only if you build it | Only through exports | Only if you build it |
| Best for | Live results pages from chat or code on one balance | Search across a set of sites you define | Daily position history for set keywords | Teams with engineers to keep it running |
Pricing
Pay as you goBilled per run from your AI/ML API balance
- No subscription
- No seat fee
- One balance for every tool
Each search is billed per run from your AI/ML API balance, with no subscription or seat fee. Our test searches each cost less than one cent. Check the current price in your dashboard before large runs.
FAQ
Is it legal to collect Google search results?
Reading public search results for research and SEO is common practice. Google's terms apply to the service, and how you store and use the data is your responsibility. Ask a lawyer about your specific case.
Can I use it as a keyword ranking API?
Yes, for a snapshot. Each run lists results in order, so you can find your domain's position for a query in a given country, language and device. Pin those settings, because results move: two of our test runs a minute apart, with different location settings, shared only 3 pages in their top 8.
Can I use it as a Google keyword position API with ranking history?
It does not store history for you. Schedule the same queries and save each result to build your own history, or use Historical Google Search Results for monthly snapshots of past result pages.
Does it include the AI Overview?
Yes, when Google shows one. You get the text and the cited sources with their links. In our run the overview cited 4 web pages and 4 product listings, and none of the 4 web pages were among the 8 organic results.
How many results can I get per query?
Up to 10 per page, then a token fetches the next page. Our run asked for 10 and got 8 organic results next to the page features, so ask for a second page when you track positions past the top few.
Does it return the text of the result pages?
It returns titles, URLs and snippets. To read a result in full, open its URL with Page Reader.
Can I use it from Claude, ChatGPT or Cursor?
Yes. Add the AI/ML API MCP server once and ask in plain English. Hermes, OpenClaw and other MCP clients work the same way.
How fresh are the results?
Each run searches Google live. Our 16 November 2026 run came back in under 9 seconds and included a video posted on 23 September 2026 at the top of the page.
Can I export the results to CSV or Google Sheets?
Ask your assistant to return the table as CSV, or write the JSON from the REST response to a file or a sheet in your code.
Run your first Google search
Connect once, then ask for any results page in plain English.