Plain-language definitions for the terms used across ScraperFleet.
A packaged automation on the Apify platform. An actor takes an input, does its work (scraping, extraction, enrichment), and writes results to a dataset.
Billing model where you are charged for the events you trigger, like one scrape call or one tool invocation, rather than for compute time.
Billing model where you are charged per result returned in the actor's dataset. Nothing is charged when a run returns nothing.
The structured output of an actor run. Datasets can be downloaded as JSON, CSV, Excel, or HTML, or read via API and webhook.
An open protocol that lets AI agents call external tools. An MCP server exposes tools an agent can invoke; ScraperFleet's MCP actors turn Apify actors into MCP servers.
Systems websites use to block automated traffic. Residential proxies and browser-based scraping are the common workarounds.
An IP address from a real consumer device. More trusted by target sites than datacenter IPs, so it is used when sites block datacenter traffic.
A cap on how many requests you can make in a time window. Good actors back off and retry instead of failing.
Turning a web page into defined fields (name, price, email) instead of raw HTML.
One execution of an actor with a given input. Every run gets an ID, status, and dataset.