Apify · Search

Apify / run any actor (submit-then-poll, async)

Runs any Apify actor and waits for it to finish. Send the actor id and the actor's own input. Returns the Apify run object (status, timing, defaultDatasetId, usageTotalUsd, stats), not the scraped rows. To get the rows, call apify-dataset-items with the run's defaultDatasetId. Set Apify's timeout parameter to bound a long run; a run that outlives the wait is not aborted.

Private Gateway connection

Data below comes from the configured private Gateway. Provider activation remains governed by its evidence and policy gates.

unverifiedRequest shape unavailableUnverified

Published API evidence

Public metadata only. Response bodies, credentials, and internal review notes are never displayed.

No verification date is claimed. No response capture is claimed.

Request parameters

input body · optional
The actor's own input document, sent as the request body exactly as Apify's Run Actor API expects (the body IS the actor's input, never wrapped in an 'input' key). Shape is actor-specific -- read it from apify-actors-builds-openapi-json for this actorId, or the actor's public Apify Store page, before calling. Optional: omitting it runs the actor with its own default input, when it has one.
maxItems query · optional
Server-side cap on results returned/produced -- always set this for a bounded call. Example: 10
maxTotalChargeUsd query · optional
Hard USD ceiling Apify enforces on this run's pay-per-event charges. Kept at round 3's bound (0.001-0.04) -- real metering (see price.estimateFormula) makes billing ACCURATE, it does not raise what's actually safe to let Apify itself spend on this shared account, which is still governed by this cap. Example: 0.04
memory query · optional
Actor run memory in MB. Example: 512
timeout query · optional
Seconds before APIFY ITSELF aborts the run server-side. Raised from round 3's sync ceiling (120) since this endpoint no longer runs inside the gateway's own 15s outbound-fetch window -- it submits and polls, so a slower cold start no longer needs to fit in 15s. Bounded 5-300 so one call still can't run forever; see the description's 'No cleanup-on-timeout' note for why this bound is the only real protection. Example: 60

Provider docs How to run it