> ## Documentation Index
> Fetch the complete documentation index at: https://docs.spidra.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Google Search

> Scrape a Google results link from any tab (web, images, news, videos, books, places) or a Google News link and get the results as structured data, or start a crawl from one and read every result.

If you already have a Google results link, paste it. Spidra reads it the way it reads its own [Search](https://docs.spidra.io/search.md): every result in order with its title, link and snippet, plus whatever else the page showed, such as an answer box, top stories or related searches. It works on every Google country domain and on the web, images, news, videos, books and places tabs. Send the link with `"output": "json"`, and each results page costs 1 credit.

To search from a query instead of a link, use [Search](https://docs.spidra.io/search.md).

## Example

```json theme={null}
{ "urls": ["https://www.google.com/search?q=best+project+management+software&gl=us&hl=en"], "output": "json", "wait": true }
```

The item's `result`, shortened to one result:

```json theme={null}
{
  "kind": "web",
  "query": "best project management software",
  "page": 1,
  "gl": "us",
  "hl": "en",
  "nextPageUrl": "https://www.google.com/search?q=best+project+management+software&hl=en&gl=us&start=10",
  "resultsCount": 9,
  "organicResults": [
    {
      "position": 2,
      "title": "Project management tools ranked + comparison table (2026 ...",
      "url": "https://www.reddit.com/r/projectmanagement/comments/1r7dr5i/project_management_tools_ranked_comparison_table/",
      "domain": "reddit.com",
      "snippet": "If you need enterprise-grade process control, Jira is top tier: configurable workflows, rich work categorization, and hierarchy (work items → ...",
      "date": "Feb 18, 2026",
      "rating": null,
      "sitelinks": []
    }
  ],
  "answerBox": null,
  "topStories": [],
  "relatedSearches": ["top 10 project management tools", "best project management software free", "best project management software for small teams", "..."],
  "serpFeatures": ["inlineImages", "relatedSearches"]
}
```

`serpFeatures` lists the blocks the page showed. `nextPageUrl` is the next results page, so you can send it in a [batch](https://docs.spidra.io/batch-scrape.md) to go deeper.

## Links that work

| Link | Example | You get |
| - | - | - |
| Web | `google.com/search?q=<words>` | Results with position, title, link, snippet, date, rating and sitelinks, and the answer box (weather, stocks, time, currency, sports, definitions), knowledge snippet, top stories, local results, images and related searches when the page has them |
| Images | `&udm=2` or `&tbm=isch` | 20 images a page with the image link, the page it's on, size and thumbnail. Google doesn't show titles here |
| News | `&udm=12` or `&tbm=nws` | Title, link, source, date, snippet and thumbnail |
| Videos | `&udm=7` or `&tbm=vid` | Title, link, platform, duration, date and thumbnail |
| Books | `&udm=36` or `&tbm=bks` | Title, Google Books link, description and cover |
| Places | `&udm=1` or `&tbm=lcl` | The places for that search, from Google Maps. See [places](#places) |
| Google News | `news.google.com` search, topic and top stories links | Up to 100 articles with title, source, exact publish time and related coverage |

Any country domain works, such as `google.co.uk` or `google.de`, and the link keeps its `hl`, `gl`, `start`, `tbs` (date filters), `safe` and location settings. Google ignores `num`, so Spidra drops it and says so in `warnings`.

A few other Google links work too: a business's reviews link and a `ludocid=` profile are read as a [Google Maps](https://docs.spidra.io/sites/google-maps.md) place, `share.google` and `g.co/kgs` short links are followed, and a `google.com/url` redirect is read as the page it points to when Spidra has a rule for that page.

## Places

The Places tab only shows a few places a page. Spidra reads the same search from Google Maps instead, which returns 20 a page with the full record, and keeps each place's rank on the tab as `tabPosition`:

```json theme={null}
{ "urls": ["https://www.google.com/search?q=coffee+shops+in+austin+tx&udm=1&gl=us&hl=en"], "output": "json", "wait": true }
```

That returned 23 places. The first, shortened:

```json theme={null}
{
  "position": 1,
  "tabPosition": 2,
  "title": "Terrible Love Coffee",
  "rating": 4.9,
  "reviewCount": 433,
  "category": "Coffee shop",
  "address": "3908 Avenue B, Austin, TX 78751",
  "website": "http://terriblelovecoffee.com/",
  "hours": "Open · Closes 2 PM",
  "hoursByDay": { "monday": "7:30 AM–2 PM", "saturday": "8:30 AM–3 PM", "...": "..." },
  "location": { "lat": 30.3045132, "lng": -97.735632 },
  "placeId": "ChIJP8p2kJC1RIYR2qrGoXZtZCk",
  "mapsUrl": "https://www.google.com/maps?cid=2982629209513831130"
}
```

Put the place in the query, as above, or add `near=<place>` to the link. Without either, Google picks local results for wherever the request came from, and `location` on the result says where that was.

## Crawl from a Google link

Start a [crawl](https://docs.spidra.io/crawl.md) with a web, news, video or image results link as its `baseUrl`, and Spidra reads the results pages, then visits each result, on whatever site it's on. Your `crawlInstruction` decides which results to keep, and a `transformInstruction` or schema runs on each one. `maxPages` caps how many results it visits.

You pay for the result pages the crawl reads, not for the Google results pages. Each kept page carries its `searchPosition`, and the crawl's [report](https://docs.spidra.io/crawl.md) lists every results page with each result's rank. A Google News link works too: each article is followed through to the publisher's own page. Books and places links can't start a crawl.

## Credits

Each results link is 1 credit on every tab. A link that can't be read is free: a search with no results, a page Google blocked, a `/search` link with no query, or a tab Spidra doesn't read.

## Limits

* People Also Ask, AI Overviews, ads, shopping results and the full knowledge panel aren't returned. The layout Spidra reads doesn't include them.
* The Shopping, AI Mode, Forums and Short videos tabs aren't read. A link to one fails free and names the tab.
* Google Maps, Jobs, Flights, Scholar and Trends links are read by their own pages: see [Google Maps](https://docs.spidra.io/sites/google-maps.md), [Google Flights](https://docs.spidra.io/sites/google-flights.md), [Google Scholar](https://docs.spidra.io/sites/google-scholar.md) and [Google Trends](https://docs.spidra.io/sites/google-trends.md).
* Google sometimes blocks a request. Spidra retries on new connections, and a page it still can't read is free.

## Templates

[Templates](https://docs.spidra.io/templates.md) for Google results include local results by keyword and place, Google News articles for a topic, image and video results, keyword ideas from related searches, the pages Google shows for a site, and reading every result on a search to pull data from each one.

> [!NOTE]
>
> Check Google's terms before you collect data from it. You're responsible for how you use what you collect.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.