Response format

The data payload of a SERP response (from live search or a completed task) uses these fields:

Field	Type	Notes
`keyword`	string	The query that was searched.
`search_engine`	`"google"`	Always `google` today.
`search_engine_domain`	string	e.g. `google.com`, `google.co.uk`.
`device`	`desktop` \| `mobile`	Device the SERP was rendered for.
`location`	`{ code, name }`	Resolved location.
`language`	`{ code, name }`	Resolved language.
`result_url`	string	The Google URL that was fetched.
`fetched_at`	string (ISO 8601)	When the SERP was captured.
`total_results_count`	integer	Google’s reported total, when present.
`pages_crawled`	integer	SERP pages fetched.
`items_count`	integer	Number of items in `items`.
`feature_types`	string[]	Distinct SERP feature types present.
`spelling`	object \| null	Spelling correction/suggestion, if any.
`items`	array	The SERP items (see below).
`tracked_rank`	object \| null	Present when `track` was requested — see rank tracking.

Items

Every entry in items is a typed SERP element discriminated by type (e.g. organic, paid, local_pack, people_also_ask). All items share a ranking triplet:

Field	Meaning
`type`	The feature type.
`rank_absolute`	Position across the whole SERP (1 = top element).
`rank_group`	Position within the item’s own type group.
`page`	Which SERP page the item appeared on.

Organic item example

{
  "type": "organic",
  "rank_group": 1,
  "rank_absolute": 3,
  "page": 1,
  "domain": "nike.com",
  "title": "Nike Running Shoes",
  "description": "Shop the latest running shoes …",
  "url": "https://www.nike.com/running",
  "breadcrumb": "nike.com › running",
  "highlighted": ["running", "shoes"],
  "rating": { "rating_type": "AggregateRating", "value": 4.6, "votes_count": 1203, "rating_max": 5 }
}

Field availability varies by type — see the full list of feature types.

Formats

The format request field controls which items appear:

advanced (default) — every parsed feature type.
standard — organic, paid, and featured-snippet items only; items are re-ranked over the filtered subset. Use this when you only care about the classic “ten blue links” plus ads.

HTML format

With format: "html", data instead contains the raw page HTML:

{
  "keyword": "coffee shops",
  "search_engine": "google",
  "search_engine_domain": "google.com",
  "device": "desktop",
  "location": { "code": 2840, "name": "United States" },
  "language": { "code": "en", "name": "English" },
  "result_url": "https://www.google.com/search?q=coffee+shops",
  "fetched_at": "2026-07-02T12:00:00.000Z",
  "pages": [ { "page": 1, "html": "<!doctype html>…" } ]
}