DataForSEO Brain

SERP API (Capability Hub)

4 min read

SERP API (Capability Hub)

The SERP API returns live and queued search-engine results pages for a keyword, location, and language across seven engines. Sits under DataForSEO Brain → Concepts.

Overview

The SERP API is DataForSEO’s search-results scraping layer. You give it a keyword, a location, and a language, and it returns the structured ranked results for that query. It spans Google, Bing, YouTube, Baidu, Yahoo, Seznam, and Naver, each under /v3/serp/{engine}/{type}/{method}. Every engine and vertical shares one request envelope and one response envelope (tasks[] → result[] → items[]), so once you learn Google Organic the other surfaces are variations on the same shape. It is the source layer behind rank tracking, SERP-feature monitoring, and AI Overview capture.

What it covers

  • Google organic (with 50+ SERP element types) and ~15 verticals (routed in cap-serp-google-verticals): AI Mode, AI Overview, Maps, Local Finder, News, Events, Images, Jobs, Autocomplete, Ads Search/Advertisers, Finance, Dataset Search/Info, Search By Image.
  • Non-Google engines (routed in cap-serp-non-google-engines): Bing, YouTube, Baidu, Yahoo, Seznam, Naver.
  • Three execution shapes per endpoint where supported: Task POST + Tasks Ready + Task GET (Standard queue), and Live (synchronous).
  • Three result depths: regular (organic + paid only), advanced (all 50+ element types), and html (raw page HTML). See cap-result-tiers-regular-advanced-html.
  • Google Organic result.items[] exposes element types including organic, paid, featured_snippet, people_also_ask, local_pack, knowledge_graph, top_stories, shopping, ai_overview, and dozens more.
  • Crawl-control inputs: max_crawl_pages, stop_crawl_on_match with match_type/match_value, target and target_search_mode, and find_targets_in/ignore_targets_in to halt early on a domain hit.
  • Helper reference endpoints for locations and languages per engine support the targeting parameters described in cap-locations-languages-targeting.

Root-level SERP utilities (engine-agnostic, run off a prior SERP task_id)

These two endpoints operate on the result of an earlier SERP task rather than on a fresh keyword query, so they live at the root of /v3/serp/ rather than under an engine.

  • AI Summary (POST /v3/serp/ai_summary, $0.01/req): generates an LLM summary over a stored SERP. Takes the source task_id (valid 30 days), a prompt (⇐ 2000 chars), support_extra (default true), fetch_content (default false), and include_links (default false). Feeds GEO and AI-search work in cap-geo-ai-search-optimization and play-ai-visibility-tracking.
  • Screenshot (POST /v3/serp/screenshot, $0.004/req): renders a full-page image of a prior SERP; the source task_id is valid for 7 days.

Key parameters / inputs

fieldnotes
keywordrequired; up to 700 chars; special operators (site:, filetype:, allinanchor) incur a 5x keyword charge
location_code / location_name / location_coordinateone location identifier required
language_code / language_namelanguage identifier
devicedesktop (default) or mobile; os follows device
depthGoogle Organic default 10, max 200
group_organic_resultsgroups related results under the main item; default true
load_async_ai_overviewfetches async AI Overview, +$0.002
people_also_ask_click_depthclicks PAA 1-4 deep, +$0.00015 per click
calculate_rectanglespixel-position ranking, +$0.002
tag / postback_url / pingback_url / prioritytask tagging, webhook delivery, queue priority

Response / what you get back

The envelope carries version, status_code, status_message, time, cost, tasks_count, tasks_error and a tasks[] array. Each task holds id, status fields, cost, result_count, path, echoed data, and result[]. Each result reports keyword, type, se_domain, check_url, datetime, item_types, se_results_count, items_count, and items[]. Organic items carry rank_group, rank_absolute, position, domain, title, url, description, plus optional rating, price, links, and rectangle (pixel box).

Cost & method notes

  • Base charge is per SERP request covering up to 10 results; depth above 10 can add cost when the engine returns more. Live Google Organic Advanced billed at $0.002 per SERP request (up to 10 results) in the cost log.
  • Standard (POST → Tasks Ready → Task GET) is cheaper than Live (instant, no GET round trip). See cap-task-vs-live-execution and cap-queue-priority-cost-model.
  • Priority is 1 (normal) or 2 (high, costs more).
  • Add-ons stack: special operators 5x, async AI Overview +0.002, PAA clicks +$0.00015/click.

When to use / how it fits

Use SERP API to track rankings, capture SERP features, and snapshot AI Overviews on a schedule. It is the spine of play-rank-tracking-pipeline and a feeder for play-local-seo-tracking and play-content-strategy-brief. Targeting parameters are detailed in cap-locations-languages-targeting. For job-to-endpoint routing see dec-which-api-for-which-job; for the live-vs-queue tradeoff see dec-live-vs-standard-vs-priority.

Gotchas / limits

  • Up to 2000 combined POST + GET calls per minute; max 100 tasks per POST; keyword ⇐ 700 chars; depth ⇐ 200; max_crawl_pages ⇐ 100.
  • Standard tasks report status 20100 “Task Created” then 40602 “Task In Queue” until ready.
  • Some engines and verticals are Standard-only (no Live) and retain task results for 30 days; this is engine-specific and detailed in the routed notes.
  • Live data is fresh-scraped; for cheaper near-real-time keyword and ranking facts, weigh pre-indexed Labs and Databases via dec-labs-vs-live-apis.

Sources

Graph View

Backlinks