--- title: "Apify Integration" description: "Run Apify Actors and saved scraping tasks from your flows. Agents launch web scrapes on demand, poll or abort runs, read dataset items and key-value store records, and append structured results for downstream steps." url: https://flowrunner.ai/integrations/apify date_modified: 2026-08-01T02:40:32-07:00 --- # Apify [Developer Tools](https://flowrunner.ai/integrations/category/developer-infrastructure) Run Apify Actors and saved scraping tasks from your flows. Agents launch web scrapes on demand, poll or abort runs, read dataset items and key-value store records, and append structured results for downstream steps. [Verified](https://flowrunner.ai/integrations/verified "What does verified mean?") · 12 actions · API key · available [Apify website](https://apify.com/) · [Platform Documentation](https://docs.apify.com/api/v2) · Capability data verified 2026-08-12 1. The Monday morning competitive price check kicks off on schedule 2. Run Task launches the saved scraping configuration for the competitor catalog 3. Get Run polls the run's status and statistics as the crawl progresses 4. Runtime, pages crawled, and dataset size are compared against the run's normal envelope 5. When a run balloons past its envelope, the operator decides whether it finishes or dies 6. Abort Run gracefully stops the crawl if that is the call, before it burns the month's compute 7. Get Dataset Items pulls the structured results from the run's default dataset 8. The pricing channel receives the diff against last week, changes highlighted ## What This Integration Enables The most current data about your market is published on the open web every day, and almost none of it arrives through an API. Apify closes that gap with Actors, reusable cloud programs for scraping and automation, and this connector makes them a managed part of your workflows rather than a script someone babysits. FlowRunner agents launch Actors and saved tasks on demand, poll runs to completion, read structured results from datasets in the format downstream steps want, pull an Actor's OUTPUT from its key-value store, and append records to datasets to aggregate results across flows. The operational stance matters more than the scraping itself: a crawl is a spend, a run has a normal shape, and results are only useful when they land where decisions happen. That is why this connector pairs naturally with FlowRunner's [human-in-the-loop](https://flowrunner.ai/concepts/human-in-the-loop) gates: agents run the collection, and a person rules on the runs that stop looking normal. ### Without FlowRunner **Scraping as a side project**: Someone's script runs on someone's laptop until the site changes or the laptop sleeps **Results trapped in files**: Scraped output lands in a CSV on a desktop, disconnected from the systems that need it **Runaway crawls found on the invoice**: A selector change turns a 10-minute scrape into an all-night crawl, discovered at billing time ### With FlowRunner **Scraping as an operation**: Actors and saved tasks run on schedule with statuses, statistics, and retries an agent can manage **Results as pipeline data**: Dataset items flow directly into sheets, stores, and downstream flows the same run **Runs watched while they run**: Agents compare live statistics against the envelope and a person rules on anything abnormal ## Use Case Scenarios ### Competitive pricing that refreshes itself A saved task holds the pricing crawl configuration. Every Monday, Run Task fires it, Get Run polls until SUCCEEDED, and Get Dataset Items returns the catalog as structured records. The agent diffs against last week's snapshot, writes the full result to [Google Sheets](https://flowrunner.ai/integrations/google-sheets), and posts only the changed SKUs to [Slack](https://flowrunner.ai/integrations/slack). Pricing reviews start from data collected that morning, not from whoever last remembered to check. ### Quick lookups inline, long crawls managed For a fast job, an enrichment lookup or a single listing page, the agent calls Run Actor And Get Items and gets the records back in the same step. For the deep crawl of a large target, it switches to the asynchronous pattern: Run Actor with input matching the Actor's schema, Get Run to poll, Get Dataset Items to collect, and the archive lands in [S3](https://flowrunner.ai/integrations/s3) for retention. One connector, two rhythms, chosen per job instead of forcing every scrape through the same timeout. ### A dataset as the team's aggregation point Signals arrive from many flows: scraped listings, form submissions, review mentions. Each flow appends its records to a shared dataset with Push Items To Dataset, which is append-only by design, so nothing overwrites history. A weekly flow reads the accumulated items, and Get Key-Value Store Record pulls each source run's OUTPUT for provenance. The team gets one queryable body of evidence instead of nine files named final\_v2. ## Human-in-Loop Highlight Abort Run is the brake this connector exists to pull, and the judgment about pulling it stays human. Actor runs consume paid compute against your Apify plan, and a runaway run is quiet: a selector change or an unexpected pagination loop turns a bounded crawl into one that grows all night, with nothing visibly wrong until the usage bill. So the agent watches Get Run statistics against the run's normal envelope, and when a crawl doubles its usual runtime or dataset size, it does not guess. It posts the run's statistics, the trend, and the projected overrun, and the operator chooses: let it finish because the target actually grew, abort gracefully so partial results survive, or force-terminate now. Killing a run also discards the completeness of its data, which is exactly the trade a person, not a threshold, should weigh. Agent processes routinely Detects exception requiring judgment Clear match Continues automatically Ambiguous Routes to human via preferred channel Human decides Agent resumes with decision ## Agent Capabilities 12 actions ### Actors 4 - **List Actors** Lists the Actors available to the account, your own or added from the Apify Store, with paging and sorting. The catalog of what your flows can run. - **Get Actor** Retrieves an Actor's default build, input schema reference, run options, and stats. Read before running something unfamiliar. - **Run Actor** Starts an Actor run asynchronously with JSON input and optional build, timeout, and memory overrides. The launch step for long crawls. - **Run Actor And Get Items** Runs an Actor synchronously and returns its dataset items inline. The right call for quick jobs that finish inside a request timeout. ### Runs 2 - **Get Run** Retrieves a run's status, timings, statistics, and its default dataset and key-value store IDs. The polling and monitoring step behind every managed crawl. - **Abort Run** Stops a running or queued run, gracefully by default or immediately with Force. The brake this page's human gate controls. ### Datasets 3 - **List Datasets** Lists the account's datasets with paging, optionally hiding the unnamed ones auto-created by runs. - **Get Dataset Items** Retrieves a dataset's records with paging, cleaning, and multiple output formats. The primary way results leave Apify and enter your flow. - **Push Items To Dataset** Appends records to a dataset. Append-only by design, which makes it a safe aggregation point across flows. ### Key-Value Stores 1 - **Get Key-Value Store Record** Retrieves a single record, such as a run's INPUT or OUTPUT, by store ID and key. The provenance read for auditing what a run was asked and what it answered. ### Tasks 2 - **List Tasks** Lists saved Actor configurations with paging. The library of repeatable scrape jobs. - **Run Task** Runs a saved task asynchronously, optionally overriding input fields for this run. The scheduled workhorse of recurring collection. ## Frequently Asked Questions ### What can FlowRunner do with Apify? FlowRunner agents can run List Actors, Get Actor, and Run Actor in Apify, plus 9 more actions. ### Does connecting Apify to FlowRunner require OAuth? No. Apify connects to FlowRunner with an API key, no OAuth flow required. ### Can Apify trigger a FlowRunner workflow automatically? Apify doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger. **Work at Apify?** This integration exposes Apify to AI agents on every FlowRunner plan, including through MCP, at no cost to you. [See what FlowRunner offers integration partners](https://flowrunner.ai/integrations/partners), including how to keep this page current. --- Markdown version of https://flowrunner.ai/integrations/apify. Site index: https://flowrunner.ai/llms.txt