The Google Search scraper by cloro is a SERP API that returns the live Google results page as parsed JSON. The JSON holds what Google shows: organic results, ads, People Also Ask, People Are Saying, local pack, related searches, shopping cards and the AI Overview.
In short, cloro's Google Search API:
- runs a new Google search for each request, so the results are not from a cached index;
- returns the raw HTML of each page beside the parsed JSON, at no extra credit cost (
include.html); - targets a country, a city (
locationoruule), a language (hl) and a device (desktop,mobile,ios,android); - returns up to 10 pages (100 results) in one response with
pages, after Google removednum=100; - has an official Python SDK, an official Node SDK and a hosted MCP server;
- runs its own proxies and automatic retries, so you solve no CAPTCHAs, and bills only successful requests.
- Get an API key at cloro.dev.
- POST a query to
https://api.cloro.dev/v1/monitor/google. - Read the parsed fields from the JSON response.
import requests
payload = {
'query': 'best serp api',
'country': 'US',
'pages': 1,
'include': {'aioverview': {'markdown': True}},
}
response = requests.post(
'https://api.cloro.dev/v1/monitor/google',
headers={'Authorization': 'Bearer YOUR_API_KEY'},
json=payload,
)
print(response.json())curl -X POST https://api.cloro.dev/v1/monitor/google \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"query": "best serp api", "country": "US", "include": {"aioverview": {"markdown": true}}}'Node.js and async/webhook examples are in the endpoint documentation.
| Parameter | Description | Default |
|---|---|---|
query* |
The search query | – |
country* |
Country code for localized results (US, GB, DE). Required unless you send gl |
– |
location |
Google canonical location name for geo-targeting. Mutually exclusive with uule |
– |
uule |
Pre-encoded Google UULE string. Mutually exclusive with location |
– |
device |
desktop, mobile, ios or android |
desktop |
hl |
Interface language, as Google's own hl (en, de, pt-BR). Overrides the language cloro derives from the country |
Derived from country |
pages |
Number of result pages to return | 1 |
include.aioverview |
Send {"markdown": true} to get the AI Overview block, with Markdown |
– |
include.html |
Return URLs to the full HTML, one per page (expire after 24h) | false |
* Required
{
"success": true,
"result": {
"organicResults": [
{ "position": 1, "title": "Best SERP APIs", "link": "https://example.com/serp-apis", "displayedLink": "https://example.com", "snippet": "Comparison of providers..." }
],
"peopleAlsoAsk": [{ "question": "What is a SERP API?", "type": "LINK", "snippet": "A SERP API returns search results as structured data...", "link": "https://example.com/what-is-a-serp-api" }],
"peopleAreSaying": [{ "position": 1, "title": "We switched after the num=100 change", "link": "https://www.reddit.com/r/SEO/comments/example/", "date": "3 days ago" }],
"localResults": [{ "position": 1, "title": "Example Agency", "rating": 4.6, "reviews": "128", "address": "123 Example St" }],
"aioverview": { "text": "SERP APIs return search results as structured JSON...", "sources": [{ "position": 1, "url": "https://example.com", "label": "Example" }] },
"relatedSearches": [{ "query": "serp api pricing", "link": "https://www.google.com/search?q=serp+api+pricing" }]
}
}organicResults: position, title, link, displayed link, snippet and "Read more" link per result.aioverview: the AI Overview block with its own cited sources, returned inline rather than behind a second request.peopleAlsoAsk: PAA questions, with the answer snippet and source link where Google shows one.peopleAreSaying: the forum and social block Google now surfaces on many commercial queries.localResults: the local pack with rating, reviews and address.shoppingCards: product carousels with price and store.ads: paid results, parsed and positioned.relatedSearches: the related-query block, each with its query and search link.- Other blocks:
knowledgeGraph,videos,shortVideos,inlineImages,hotelResults,flights,transitResults,topSights,events,stores,comparisonShoppingSites,weather,sportsResults,financeQuote,discussionsAndForums,latestPosts,thingsToKnowandrelatedWebsites, each described in the endpoint reference.
Google shows these blocks only on some queries, and a key is absent when its block is not on the page.
Full field-level schemas are in the endpoint reference.
- Rank tracking at city level using
locationoruule, where a national average and a metro result diverge. - AI Overview monitoring: presence, text and cited sources, in the same response as the organic results.
- SERP feature research: how often PAA, local pack or People Are Saying fire on a query class.
- Competitive analysis: who holds the page for the queries your buyers run.
cloro has both. The Python SDK is on PyPI as cloro (pip install cloro, source in cloro-python), and the Node SDK is @cloro-dev/cloro on npm. The hosted MCP server is at https://mcp.cloro.dev/mcp: Claude connects with OAuth, and Cursor or any other MCP client connects with an API key (MCP setup). The MCP tool for this endpoint is scrape_google. In Python, client.monitor.google(query=..., country=..., pages=...) calls this endpoint.
Live results. cloro runs a new Google search when you send the request, and parses the page that Google returned (how cloro retrieves data). cloro keeps no index of its own. Send include.html: true to get the page Google served as proof.
Google removed num=100 on September 11, 2025, so every provider now paginates at 10 results a page. With cloro, set pages from 1 to 10. Each page holds 10 results, and all pages come back in one response. Page 1 costs 3 credits async, and each extra page costs 2 credits more, so 100 results cost 21 credits async, or 23 credits as a sync request (pricing).
You need no proxies with cloro, and you solve no CAPTCHAs. cloro sends each request through its own proxy in the country you set and retries a failed or blocked attempt automatically, up to 5 attempts (retries). A blocked attempt costs nothing.
cloro's rate limit is 1,000 requests per second per endpoint and API key. The real limit for sync requests is your plan's concurrency: the number of requests that can run at the same time, from 10 on Lite to 100 on Business and more on Enterprise plans (concurrency). For large volumes, use async tasks: one call submits up to 500 tasks, your organization can queue up to 100,000, and concurrency never blocks a submission.
cloro has no BigQuery connector, so you write the rows yourself. On your schedule (cron, Cloud Scheduler or Airflow), submit the keywords as an async batch to cloro with a webhook.url. When each task completes, cloro sends the full JSON to the webhook, and your handler writes one row per organic result to BigQuery (async requests, webhooks).
cloro documents an average success rate above 99% (error handling). Google keeps changing its pages and blocking, so the rate moves; the status page shows live numbers. You run no proxies: cloro makes up to 5 attempts per request. Background: Google's verification page and SERP API reliability at scale.
Yes: cloro's Google Search API returns it when you set include.html: true. result.html then comes back in the same response as the parsed fields, as an array of CDN URLs, one per results page, at no extra credit cost. Each URL expires after 24 hours, so download the HTML if you need it longer (output formats). Use it as proof of the page Google served, or to parse a block that cloro does not parse.
cloro returns a SERP with no personalization from a user account: each request runs in a logged-out session, with no account of any kind, through a proxy in the country you set (how cloro retrieves data). The page still changes with the inputs you send, by design. location or uule sets the place (without either, cloro uses the country's canonical location where it has one), device selects the desktop or mobile SERP, and hl sets the interface language. Send the same inputs on every run when you compare results over time. To match what one user sees, send that user's country, location, hl and device; a logged-in user can still see personalized results.
Yes. It is a paid add-on, charged whether or not Google shows an Overview for the query, and it comes back in the same response rather than a second request. See Google Search pricing for how credits add up with pages.
Yes. For a city, location takes a Google canonical location name and uule takes a pre-encoded string. They are mutually exclusive. For a device, set device to desktop, mobile, ios or android.
Yes. cloro charges a sync request when it returns 200 and an async task when it reaches COMPLETED. Errors, failed tasks and the retries cloro runs inside a request cost nothing (billing).
Yes. The n8n-nodes-cloro community node is listed in the n8n integrations directory, and the cloro app for Zapier sends a query from any Zap. Both return the parsed JSON.
cloro returns publicly visible results pages. No court has ruled that scraping public results is unlawful, though the area is active. Check your own jurisdiction and use case.
- Endpoint reference: cloro.dev/docs
- Product page: cloro.dev/google-search
AI Mode · AI Overview · Bing AI Mode · Bing Search · ChatGPT · Gemini · Google News
Questions or support: ask the docs AI assistant.
