DataForSEO Brain

Task (Standard) vs Live Execution

4 min read

Task (Standard) vs Live Execution

Two ways to run any DataForSEO job: the asynchronous Standard queue (task_post → tasks_ready → task_get) or synchronous Live that returns results in the same request. Sits under DataForSEO Brain → Concepts.

Overview

Every applicable module exposes one or both of two execution modes. Standard (task-based) is asynchronous and uses three separate calls; Live is synchronous and returns data inside one open connection. The choice drives latency, cost, throughput, retry logic, and whether results are stored. This is a hub note: most module notes reference it to explain how their endpoints are invoked.

What it covers

  • Standard method: task_post (submit), tasks_ready (list completed-but-uncollected task IDs), task_get (retrieve results). Described as “more affordable” and preferred for planned, scalable collection.
  • Live method: synchronous, no separate POST/GET; the connection stays open until data is ready (“on average it takes up to 6 seconds,” longer for large queries). Most expensive mode.
  • Mode availability by API: SERP, Merchant, and Business Data support both Live and Standard (with priority options); DataForSEO Labs is Live-only; OnPage is Standard-based with cost/speed set by crawl parameters.
  • Collection without polling: register pingback_url or postback_url so completion is pushed to you. See cap-webhooks-pingback-postback.

Key parameters / inputs

  • Standard POST body: array of up to 100 task objects; each can carry tag, pingback_url, postback_url, postback_data (“advanced”/“html”/“regular”).
  • Live POST: 1 task per request (Live methods generally allow only one task/POST).
  • priority: 1 (Normal) or 2 (High) on Standard only - trades latency for cost.
  • task_get tail selects result depth: .../task_get/advanced/$id or .../task_get/regular/$id or .../task_get/html/$id. See cap-result-tiers-regular-advanced-html.

Response / what you get back

  • task_post returns 20100 (“Task Created.”) with a per-task id (UUID) you later poll/collect; no result payload yet.
  • tasks_ready returns the IDs of finished tasks not yet collected.
  • task_get and Live return the full result array in the standard envelope (status_code 20000 on success).
  • Live returns result immediately; nothing is queued or stored.

Cost & method notes

Live costs roughly 3.3x Standard-Normal and ~1.65x Standard-High for the same SERP (0.0006 vs 0.002; Labs Live keyword_ideas = $0.0105. Cost detail lives in cap-queue-priority-cost-model.

When to use / how it fits

Gotchas / limits

  • Live timeout is 120 seconds (status 50401 on overrun); set your client timeout to 120s so requests aren’t cut early. Target page response cap is 50s (50402).

  • Live results are not stored - capture them from the response; Standard results are stored 30 days and re-collectable free.

  • A task returning an error is not processed further; you must re-set it after a delay. Retryable: 40103, 40601, 40602, 50301, 50303, 50401, 50402. See cap-status-error-codes.

  • Labs being Live-only means you cannot queue it; budget for synchronous calls. See cap-labs-keyword-research.

  • tasks_ready lists only completed-but-uncollected tasks; once you task_get a task it drops off the list.

  • OnPage is Standard-based; its cost/speed is set by crawl parameters rather than a Live/queue toggle. See cap-onpage-api.

  • Merchant and Business Data offer both modes with priority options. See cap-merchant-api and cap-business-data-api.

  • Labs is Live-only, so every Labs call counts against the 30-simultaneous concurrency cap. See cap-rate-limits-throughput.

  • Live averages up to 6 seconds but large/complex queries take longer; budget for the 120-second ceiling.

  • For more than 100,000 tasks/day, use Standard plus postback/pingback rather than Live. See cap-webhooks-pingback-postback.

  • task_get supports regular/advanced/html depth tails to control parse richness. See cap-result-tiers-regular-advanced-html.

  • Spend per mode is monitorable via appendix/user_data. See cap-account-usage-userdata.

  • Test both modes for free in the sandbox before billed runs. See cap-sandbox-testing.

  • Mode choice is the core of the routing in dec-which-api-for-which-job.

  • DataForSEO API v2 was fully closed on 2026-05-05: no new v2 requests are accepted (only legacy task_get stayed active during the sunset window), so the v3 Standard/Live split documented here is the only live option. The AI Optimization API is v3-exclusive. See cap-ai-optimization-api.

Sources

Graph View

Backlinks