Apify n8n integrationAutomate Apify with n8n.

Starts an Actor and returns the run, without necessarily waiting for it. The fire-and-forget half of the node, paired with the trigger for long crawls.

• Actor: required, in the format username~actor-name or as a plain Actor ID.
• Input JSON: overrides the prefilled input. For web scrapers, common fields are startUrls, maxCrawlPages and proxyConfiguration; pass {} for the Actor defaults.
• Wait for Finish: holds the execution until the run ends, capped by your workflow timeout setting.
• Maximum Cost per Run (USD): aborts the run once the amount is reached, on Actors priced per event or per result.

One node, three steps: it starts the Actor, waits for it to finish, and returns the dataset items. The rows arrive on the output, so the next node works on real data instead of a run ID.

• Actor: required, the Actor to run, browsable in Apify Store.
• Input JSON: the run input, whose structure depends on the Actor. The exact schema lives in Apify Console.
• Timeout: a run timeout in seconds, replacing the Actor's default run configuration.
• Memory: the memory limit in megabytes, from 128 up to 32768.

Runs one page through the Website Content Crawler Actor and gives back its content as text, markdown and HTML. No Actor ID to look up, which makes it the quickest way to read a page from a workflow.

• URL: the address to scrape. It has to start with http:// or https:// and be a valid URL.
• Crawler Type: cheerio, jsdom, playwright:adaptive or playwright:firefox, depending on how much the page relies on scripts.
• Output Format: markdown for AI agents and language models, html for programmatic processing, text for plain content.

Fetches the most recent run of an Actor, with its status and its storages, without you having to know a run ID first. Handy when the workflow starts somewhere other than Apify.

• Actor: required, the Actor ID or a tilde-separated username and Actor name.
• Status: narrows the answer to runs in one state, among READY, RUNNING, SUCCEEDED, FAILED, ABORTING, ABORTED, TIMING-OUT and TIMED-OUT.

Tasks are Actor configurations saved in Apify Console, so this operation runs one by name and returns all its details. The input lives on the Apify side, which keeps the n8n node short.

• Actor Task: required, as username~task-name or a plain task ID.
• Use Custom Body: switch it on to override the input saved in the task.
• Input (JSON): the override itself, following the input schema of the Actor behind the task.
• Build Tag: the build to run, instead of the one set in the task settings.

Same saved task, but the node waits and hands back the dataset items directly. It removes the second hop when the task finishes quickly enough to sit inside one execution.

• Actor Task: required, the task to run, managed in Apify Console.
• Use Custom Body: whether the saved input gets replaced for this run.
• Timeout: a run timeout in seconds, overriding the task settings.
• Maximum Cost per Run (USD): the ceiling in USD past which the run is aborted. Leave it empty for no limit.

Returns the run history of the whole account, not of one Actor. It is the operation behind any reporting on what actually ran, and it takes no required field at all.

• Limit: how many runs come back at most.
• Offset: how many entries to skip at the start, 0 by default, which is how you walk through a long history.
• Desc: flips the startedAt sort to descending, so the latest runs come first.
• Status: keeps only runs in one terminal state, among SUCCEEDED, FAILED, TIMED-OUT and ABORTED.

Looks up one run by its ID and returns its metadata, its status and its storages. Most workflows use it right after a trigger event, to turn a bare run reference into something worth reading.

• Run ID: required, and typically an expression such as {{ $json.id }} pointing at the run that just ended. A required field left empty stops the node from executing.

Lists the runs of one specific Actor, which is the difference with the account-wide history. Use it to look at a single crawler over time.

• Actor: required, the Actor whose history you want, as an ID or a tilde-separated username and Actor name.
• Status: filters the list, and here an empty value stands for All alongside the individual states.
• Desc: reverses the default ascending sort on startedAt.
• Limit: the maximum number of results, paired with Offset to page through them.

Stops a run that is still going. The only operation on this node that changes something rather than reading it, which is worth remembering when you wire it behind anything automatic.

• Run ID: required, the ID of the run to abort, usually carried in from a previous node.
• Graceful: gives the Actor time to finish its current task before it stops. Switched off, the run is killed immediately.

Reads the rows out of a dataset, which is where an Actor puts what it collected. This is the operation that turns a finished run into items your other nodes can work on.

• Dataset ID: required, either the ID or username~dataset-name.
• Limit: the maximum number of items returned, with Offset skipping the first ones so a large dataset comes back in slices.
• Fields: a comma-separated list of the only fields to keep, which keeps wide scraping results manageable.
• Omit: the mirror image, dropping fields; a field in both lists is excluded.

Key-value stores hold the single artefacts a run produces, a screenshot or a summary file rather than a table of rows. This operation reads one of them by key.

• Key-Value Store ID: required, the store to read from, usually the one attached to the run you just fetched.
• Key-Value Store Record Key: required, the key of the record itself. Both are resource locators, so you can pick from a list or set them with an expression.