SERP API

The Search endpoint performs a Google search for any query and returns fully parsed, structured JSON data. You get organic results, ads, knowledge graphs, local packs, video results, related questions, and more. All extracted and organized without any HTML parsing on your end.

Endpoint

GET https://api.scrape.do/plugin/google/search

Request Parameters

Required

General

Localization & Geo-targeting

Result Filtering

Example Usage

Step 1: Define Your Search Query

Decide what you want to search for on Google. This can be:

• General queries: pizza new york, best laptop 2026
• Local searches: restaurants near me, hotels in istanbul
• Informational queries: how to learn python, what is web scraping

For this example, we'll search for "pizza new york":

Step 2: Send the API Request

curl --location --request GET 'https://api.scrape.do/plugin/google/search?token=<SDO-token>&q=pizza+new+york'

import requests
import json

token = "<SDO-token>"
query = "pizza+new+york"

url = f"https://api.scrape.do/plugin/google/search?token={token}&q={query}"

response = requests.request("GET", url)

print(json.dumps(response.json(), indent=2))

const axios = require('axios');

const token = "<SDO-token>";
const query = "pizza+new+york";

const url = `https://api.scrape.do/plugin/google/search?token=${token}&q=${query}`;

axios.get(url)
  .then(response => {
    console.log(JSON.stringify(response.data, null, 2));
  })
  .catch(error => {
    console.error(error);
  });

package main

import (
	"fmt"
	"io/ioutil"
	"net/http"
)

func main() {
	token := "<SDO-token>"
	query := "pizza+new+york"

	url := fmt.Sprintf(
		"https://api.scrape.do/plugin/google/search?token=%s&q=%s",
		token, query,
	)

	resp, err := http.Get(url)
	if err != nil {
		panic(err)
	}
	defer resp.Body.Close()

	body, _ := ioutil.ReadAll(resp.Body)
	fmt.Println(string(body))
}

require 'net/http'
require 'json'

token = "<SDO-token>"
query = "pizza+new+york"

url = URI("https://api.scrape.do/plugin/google/search?token=#{token}&q=#{query}")

response = Net::HTTP.get(url)

puts JSON.pretty_generate(JSON.parse(response))

import java.net.HttpURLConnection;
import java.net.URL;
import java.io.BufferedReader;
import java.io.InputStreamReader;

public class GoogleSearch {
    public static void main(String[] args) throws Exception {
        String token = "<SDO-token>";
        String query = "pizza+new+york";

        String url = String.format(
            "https://api.scrape.do/plugin/google/search?token=%s&q=%s",
            token, query
        );

        HttpURLConnection conn = (HttpURLConnection) new URL(url).openConnection();
        conn.setRequestMethod("GET");

        BufferedReader reader = new BufferedReader(
            new InputStreamReader(conn.getInputStream())
        );
        String line;
        StringBuilder response = new StringBuilder();
        while ((line = reader.readLine()) != null) {
            response.append(line);
        }
        reader.close();

        System.out.println(response.toString());
    }
}

using System;
using System.Net.Http;
using System.Threading.Tasks;

class Program
{
    static async Task Main()
    {
        string token = "<SDO-token>";
        string query = "pizza+new+york";

        string url = $"https://api.scrape.do/plugin/google/search?token={token}&q={query}";

        using HttpClient client = new HttpClient();
        string response = await client.GetStringAsync(url);

        Console.WriteLine(response);
    }
}

<?php
$token = "<SDO-token>";
$query = "pizza+new+york";

$url = "https://api.scrape.do/plugin/google/search?token={$token}&q={$query}";

$ch = curl_init();
curl_setopt($ch, CURLOPT_URL, $url);
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);

$response = curl_exec($ch);
curl_close($ch);

echo json_encode(json_decode($response), JSON_PRETTY_PRINT);
?>

Step 3: Receive Structured Results

The API returns a JSON object with all SERP elements parsed and categorized:

{
  "search_parameters": {
    "q": "pizza new york",
    "hl": "en",
    "gl": "us",
    "device": "desktop",
    "start": 0,
    "google_domain": "google.com"
  },
  "search_information": {
    "page_title": "pizza new york - Google Search",
    "query_displayed": "pizza new york",
    "total_results": 349000000,
    "time_taken_displayed": 0.52,
    "organic_results_state": "Results for exact spelling",
    "results_for": "",
    "country": "",
    "city": ""
  },
  "organic_results": [
    {
      "position": 1,
      "title": "The Best Pizzerias of New York City",
      "link": "https://ny.eater.com/maps/best-pizza-new-york-city-nyc-pizzerias",
      "displayed_link": "https://ny.eater.com › maps › best-pizza-new-york-city-...",
      "snippet": "Where to find the best pizza in New York at the best pizzerias in NYC serving the best pies and slices including L&B Spumoni Gardens, L'Industrie, ...",
      "snippet_highlighted_words": ["best"],
      "source": "Eater New York",
      "sitelinks": {
        "expanded": [
          { "title": "L&B Spumoni Gardens", "link": "https://ny.eater.com/maps/best-pizza-new-york-city-nyc-pizzerias#lb-spumoni-gardens" },
          { "title": "L'Industrie Pizzeria", "link": "https://ny.eater.com/maps/best-pizza-new-york-city-nyc-pizzerias#lindustrie" }
        ]
      }
    },
    {
      "position": 2,
      "title": "Ny Pizza New York, NY - Last Updated February 2026",
      "link": "https://www.yelp.com/search",
      "displayed_link": "https://www.yelp.com › Restaurants",
      "snippet": "Top 10 Best Ny Pizza in New York, NY - Yelp - NY Pizza Suprema, Juliana's, L'industrie Pizzeria, Joe's Pizza, Angelo's Coal Oven ...",
      "snippet_highlighted_words": ["NY Pizza Suprema"],
      "source": "yelp.com",
      "rich_snippet": {
        "bottom": {
          "extensions": ["(41,361)"]
        }
      }
    },
    {
      "position": 3,
      "title": "NYC: Best Pizza",
      "link": "https://www.the-carboholic.com/nyc-guides/nyc-best-pizza",
      "displayed_link": "https://www.the-carboholic.com › nyc-guides › nyc-best-...",
      "snippet": "Nov 23, 2025 — The best (and one of the only) I've found is from Emmett's. You can build your own deep dish pie here, choosing from a list of 20+ toppings.",
      "snippet_highlighted_words": ["Emmett's"],
      "source": "The Carboholic"
    }
  ],
  "related_questions": [
    { "question": "What is the most famous pizza in New York?" },
    { "question": "What is the best pizza in NYC?" },
    { "question": "Why is Joe's pizza so famous?" }
  ],
  "related_searches": [
    { "query": "best pizza in new york", "link": "https://www.google.com/search?q=best+pizza+in+new+york" },
    { "query": "pizza delivery new york", "link": "https://www.google.com/search?q=pizza+delivery+new+york" }
  ],
  "pagination": {
    "current": 0,
    "next": "https://www.google.com/search?q=pizza+new+york&start=10",
    "other_pages": {
      "2": "...&start=10",
      "3": "...&start=20"
    }
  }
}

Pagination

To get additional pages of results, increment the start parameter by 10. Offset can go up to start=100; a higher value is rejected with 400 because no results exist that deep. On the last servable page, pagination.next is omitted.

Send start from the very first request of a sweep — begin at start=0 and send it on every page. Pages of the same search (identical parameters except start) are then served from one browsing session for a short window (about 2 minutes), so results stay coherent across start=0,10,20,… instead of reshuffling or repeating between pages. There is nothing to enable: sending start is the signal.

/plugin/google/search?token=...&q=pizza+new+york&start=0
/plugin/google/search?token=...&q=pizza+new+york&start=10
/plugin/google/search?token=...&q=pizza+new+york&start=20

• Omitting start on the first request and adding it only from the second page onward breaks that coherence: the first request is treated as a standalone search and can return a different set of results than the rest of the sweep.
• Send pages one at a time: request the next page only after the previous one has returned. While a page of the same search is still in progress, another page is rejected with 429 and is not charged. Pages from different searches can run in parallel. This limit applies per account, so using another API token from the same account does not bypass it.
• Allow at least 60 seconds per request. If your client stops waiting sooner, the request may still be running on our side, and an immediate retry can receive 429 until it finishes.
• Changing any other parameter that affects results (q, gl, hl, location, device, …) starts a separate search. cache, include_html and resolveGoto do not, so a page sent with cache=false still belongs to the same sweep.
• A sweep that takes longer than the window simply continues on a new session.
• A single search sent without start is a one-off and is not tied to any sweep.
• Pages after the first (start > 0) contain web results: organic_results, pagination, and, when Google shows them, top_ads and bottom_ads are filled as usual. related_questions, top_stories, local_results, and local_map are empty on these pages, while related_searches and video_results can still appear. The first page (start=0) carries every section.
• With device=mobile, Google does not return numbered pagination links, so pagination.next is omitted and pagination.other_pages is empty. Manual pagination still works: send start=0,10,20,… yourself rather than following pagination.next.

Caching

Successful searches are cached for 6 hours. Repeating an identical search during that time returns the cached page, so the results and total_results remain consistent. Cached responses cost the same credits as fresh responses.

A cached page is reused only when every result-affecting parameter matches. include_html and resolveGoto do not affect cache matching. Paginated searches are cached page by page, and pages in the same cached sweep stay consistent with each other. To refresh a complete paginated search, request the first page with start=0&cache=false; later pages then follow that fresh sweep.

Set cache=false to skip the cache and fetch a fresh page. The fresh page replaces the cached version for later requests. The previous no_cache=true parameter remains supported as an alias and has the same effect.

Pages without organic results and pages with a deferred AI Overview are not cached.

Response Structure

Top-Level Fields

search_parameters

Echoes the request parameters. Always-present fields are shown below; optional fields appear only when sent by the client.

{
  "q": "pizza new york",
  "hl": "en",
  "gl": "us",
  "device": "desktop",
  "start": 0,
  "google_domain": "google.com",
  "location": "New York,New York,United States",
  "uule": "w+CAIQICIfTmV3IFlvcmssTmV3...",
  "nfpr": true,
  "lr": "lang_en",
  "cr": "countryUS",
  "safe": "active",
  "time_period": "last_week",
  "filter": "0"
}

search_information

Search metadata. All fields are always present (empty string or zero when no data).

organic_results[]

Standard web search results.

top_ads[] & bottom_ads[]

Sponsored results. top_ads appear above organic results, bottom_ads appear below.

knowledge_graph

Rich knowledge panel. Fields vary by entity type (person, business, place, etc.).

local_results

Local business results (Google Maps pack). Contains a places array and an optional link to more results.

{
  "places": [
    { "position": 1, "title": "Joe's Pizza Broadway", "rating": 4.5, "price": "$10–20" },
    { "position": 2, "title": "NY Pizza Suprema", "rating": 4.6, "price": "$10–20" }
  ],
  "more_locations_link": "https://..."
}

local_map

Map widget associated with local results.

{
  "link": "https://...",
  "image": "data:image/png;base64,..."
}

related_questions[]

"People Also Ask" expandable question boxes.

video_results[]

top_stories[]

News and stories carousel. Contains sections with an array of story items.

[{
  "items": [
    { "position": 1, "title": "...", "link": "..." },
    { "position": 2, "title": "...", "link": "..." }
  ]
}]

discussions_and_forums[]

related_searches[]

ai_overview

When Google returns an AI Overview for the query, this field is included in the first page (start=0). Pages after the first return web results without an AI Overview. When no AI Overview is present, the field is null.

AI Overview uses these response states:

Complete example:

{
  "ai_overview": {
    "state": "complete",
    "text_blocks": [
      {
        "type": "paragraph",
        "snippet": "To print a document, open the file, select File > Print or press Ctrl+P (Windows) / Cmd+P (Mac)...",
        "reference_indexes": [0, 1, 2]
      },
      {
        "type": "list",
        "list": [
          {
            "snippet": "Open the File: Open the document you wish to print."
          },
          {
            "snippet": "Access Print Menu: Click \"File\" then select \"Print.\""
          },
          {
            "snippet": "Select Printer & Settings: Choose the correct printer.",
            "list": [
              { "snippet": "Printer: Choose the correct printer from the dropdown menu." },
              { "snippet": "Pages: Select \"All Pages\" or enter specific page numbers." }
            ]
          }
        ],
        "reference_indexes": [3, 6]
      }
    ],
    "references": [
      {
        "title": "How to Print Documents from Computer",
        "link": "https://www.youtube.com/watch?v=example",
        "snippet": "To print from a laptop, ensure both the printer and laptop are connected...",
        "source": "YouTube"
      }
    ]
  }
}

Deferred example:

{
  "ai_overview": {
    "state": "deferred",
    "session_key": "ce5a60050323b4beee15b60e1fcc9018"
  }
}

AI Overview Schema

AIOverview Object

TextBlock Object

ListItem Object

Reference Object

SnippetLink Object

Fetching a Deferred AI Overview

When you receive state: "deferred", make a follow-up request to fetch the complete AI Overview:

GET https://api.scrape.do/plugin/google/search/ai-overview?token=TOKEN&session_key=SESSION_KEY

The response returns the same AI Overview structure with state: "complete" or state: "unavailable":

{
  "state": "complete",
  "text_blocks": [
    {
      "type": "paragraph",
      "snippet": "mRNA vaccines work by instructing cells to produce a harmless piece of the target virus's spike protein...",
      "reference_indexes": [0, 1]
    }
  ],
  "references": [
    {
      "title": "Understanding mRNA COVID-19 Vaccines",
      "link": "https://www.cdc.gov/...",
      "snippet": "mRNA vaccines teach our cells how to make a protein...",
      "source": "CDC"
    }
  ]
}

Included HTML in the follow-up response

By default, html is just the HTML the AI Overview was extracted from. If you also passed include_html=true on the original search request that produced this session_key, html is instead one combined page with the AI Overview content already inserted into the original page — so you don't have to stitch the two responses together yourself. If the original page's HTML isn't available (for example, include_html=true wasn't passed on the original request, or too much time passed between the two requests), html falls back to just the AI Overview fragment.

In both cases html is ready-to-render markup: the AI Overview arrives from Google in an incremental transfer format, and it is assembled into finished HTML for you, with the answer in its normal place on the page. You can save the response to a file and open it in a browser as-is.

Result links

link is an absolute URL you can request directly — no host to prepend, no path to resolve.

• Results pointing at one of Google's own surfaces (Flights, Books, Arts, Finance) keep that host, for example https://www.google.com/travel/flights/flights-to-tokyo.html. They are ordinary results and count toward position.
• Google serves result links as its own redirect, https://www.google.com/goto?url=…, in place of the destination URL. When the page carries the destination as well, link is resolved back to it at no extra cost, so it is the publisher's URL.
• For links whose destination the page does not carry — commonly AI Overview references, image-led cards and the complaint links in dmca_messages — link is Google's redirect on the domain you searched. It lands on the destination when you follow it, or pass resolveGoto=true to get the destination in the response.

ai_overview.references[].link follows the same rules.

Resolving redirect links (resolveGoto)

With resolveGoto=true, every Google redirect link still in the response is replaced with the destination URL it points to before the response is returned:

/plugin/google/search?token=...&q=what+is+seo&gl=us&hl=en&resolveGoto=true

• All redirects in the response are resolved at the same time. A link that cannot be resolved keeps its redirect URL, and the request still succeeds.
• The destination URL replaces the redirect in the same field.
• Only link values change: the response has the same fields, results, order and position with or without resolveGoto.
• Resolving adds to the response time. Allow several extra seconds on top of the normal search latency.
• Cost: no extra credits — a request costs the same with or without resolveGoto.

dmca_messages

When Google has removed pages from a result page, it prints a notice under the results saying so. dmca_messages carries that block. The key is absent when the page shows no notice, so its presence is itself the signal that something was removed for this query.

{
  "dmca_messages": {
    "title": "Notices about Filtered Results",
    "messages": [
      {
        "content": "In response to multiple complaints we received under the US Digital Millennium Copyright Act, we have removed 2 results from this page. If you wish, you may read the DMCA complaints that caused the removals at LumenDatabase.org: Complaint, Complaint.",
        "highlighted_words": [
          { "text": "US Digital Millennium Copyright Act", "link": "https://www.google.com/support/answer/1386831" },
          { "text": "Complaint", "link": "https://www.google.com/goto?url=CAESYgHrOzAVyuo5c-Uj…" },
          { "text": "Complaint", "link": "https://www.google.com/goto?url=CAESYgHrOzAVaXZLynLN…" }
        ]
      }
    ]
  }
}

navigation[]

pagination

{
  "current": 0,
  "next": "https://www.google.com/search?q=...&start=10",
  "other_pages": {
    "2": "...&start=10",
    "3": "...&start=20"
  }
}

Example Requests

Basic search:

/plugin/google/search?token=TOKEN&q=pizza+new+york

Turkish results from Turkey, mobile device:

/plugin/google/search?token=TOKEN&q=pizza&hl=tr&gl=tr&device=mobile&google_domain=google.com.tr

Location-based search (auto UULE encoding):

/plugin/google/search?token=TOKEN&q=restaurants&gl=tr&location=Istanbul,Istanbul,Turkey

Advanced filtering:

/plugin/google/search?token=TOKEN&q=python+tutorial&lr=lang_en&cr=countryUS&nfpr=true

Pagination (2nd page):

/plugin/google/search?token=TOKEN&q=pizza+new+york&start=10

Resolve Google redirect links:

/plugin/google/search?token=TOKEN&q=what+is+seo&gl=us&hl=en&resolveGoto=true

Fetch fresh results:

/plugin/google/search?token=TOKEN&q=test&cache=false

Debug mode (include raw HTML):

/plugin/google/search?token=TOKEN&q=test&include_html=true