Glossary

Short definitions of the terms used across Apify docs. Each entry links to the page that covers the concept in full.

This page covers only what is specific to Apify, not the general vocabulary of web scraping and software development.

Actor

An Actor is a serverless cloud program that takes a structured JSON input, performs a task, and optionally produces a structured output. Actors do jobs on the web, such as data extraction, lead generation, social media monitoring, or competitor tracking, and their output equips AI agents and apps with real-time web data. You build an Actor from your own code, run it from Apify Console, the Apify API, or a schedule, and can publish it to Apify Store to earn from others using it.

Actor run

An Actor run is a single execution of an Actor, with its own container, allocated resources, and storage. Each run gets a dataset, a key-value store, and a request queue, and moves from ready to running before it ends as succeeded, failed, timed out, or aborted.

Actor Standby

Actor Standby is a mode that keeps an Actor ready in the background as an HTTP server. It answers requests in real time instead of paying a cold start on every call, which suits Actors used as live endpoints rather than batch jobs.

Actor task

An Actor task is a saved, reusable input configuration for a specific Actor. It lets you rerun the same Actor with preset inputs from a schedule, the Apify API, or Apify Console.

Agentic payments

Agentic payments let AI agents pay for and run Actors without an Apify account or ongoing billing. With x402 (stablecoins on Base or Solana) or the Machine Payments Protocol (MPP) (stablecoins on Tempo or Solana), the agent pays once for a prepaid API token through Apify AGI, Apify's entry point for agents, and runs Actors until its balance runs out. A prepaid token runs only Actors that use limited permissions. With Skyfire, the agent pays for each run with a pre-funded Skyfire token.

API token

An API token is a secret key tied to your Apify account that authenticates requests to the Apify API, the Apify CLI, the client libraries, and integrations. It carries your account's permissions, or a limited subset of them.

Apify AGI

Apify AGI (Agent General Interface) is the entry point for autonomous agents to use and pay for the Apify platform and Actors. Agents reach it through agentic payments, so they can run Actors without account setup or ongoing billing.

Apify AI

Apify AI is a conversational interface inside Apify Console for finding and running Actors in natural language, instead of keyword search. It runs on the same search and execution backend as Apify Store search and the Apify MCP server.

Apify API

The Apify API is the REST API for running Actors, managing storage, and controlling the Apify platform programmatically. It's the basis for the client libraries, the CLI, and most integrations, and authenticates with an API token.

Apify CLI

The Apify CLI is the command-line tool for creating Actors from templates, running them locally against local storage, and deploying them to the Apify platform.

Apify Console

Apify Console is the web application for managing your Apify projects and resources: you run Actors, inspect storage, and handle billing. It renders each Actor's input form from its input schema, and presents run results according to the output schema.

Apify MCP Proxy

The Apify MCP Proxy is the component of the Apify platform that an Actor reaches its MCP connectors through. It validates each request, injects your stored credentials server-side, and forwards it to the upstream MCP server, so the credentials never enter the Actor.

Apify MCP server

The Apify MCP server is the server that exposes Actors as tools over the Model Context Protocol, so AI assistants and agents can discover and run them. Clients such as Claude, ChatGPT, and Cursor connect to it and call Actors as part of a conversation, without custom integration code.

Apify Proxy

Apify Proxy is the proxy service that routes requests through datacenter, residential, and Google SERP proxies to reduce blocking. It rotates IP addresses automatically and lets you target specific countries.

Apify SDK

The Apify SDK is the JavaScript and Python toolkit for building Actors and working with the Apify platform programmatically. It wraps storage, proxy, and run lifecycle handling, so an Actor reads input and writes results without calling the API directly.

Apify Store

Apify Store is the public catalog of Actors that anyone can find and run, positioned as the marketplace of tools for AI. Anyone can build and publish Actors there and earn from others using them, with placement influenced by each Actor's quality score.

Apify storage

Apify storage is the built-in storage for Actor data, made up of datasets, key-value stores, and request queues. Every Actor run gets one of each by default, and you can also create named stores that persist across runs.

Build

A build is a versioned Docker image compiled from an Actor's source code. Builds are numbered and tagged, so a run can pin a specific version.

Compute unit

A compute unit (CU) is the Apify platform's usage and billing unit, equal to 1 GB of memory running for one hour. Allocating more memory to an Actor run consumes CUs faster.

Crawlee

Crawlee is Apify's open-source library for building web scrapers and browser automation in Node.js and Python. It handles crawling concerns such as queueing, retries, session management, and browser fingerprinting, and runs anywhere, not only on the Apify platform.

Datacenter proxy

A datacenter proxy routes requests through datacenter IP addresses. It's fast and cheap, at a higher risk of blocking, and suits high-volume work against targets without strong anti-bot protection. Addresses come from shared pools or from dedicated groups assigned only to you.

Dataset

A dataset is append-only, table-like storage for an Actor's structured results. You export it as JSON, CSV, or Excel, read it through the Apify API, and can describe its shape with a dataset schema.

Google SERP proxy

A Google SERP proxy is a proxy for collecting localized Google search engine results pages. It targets specific country domains and languages.

Input schema

An input schema is the JSON definition of an Actor's inputs. It generates the Actor's input form in Apify Console, validates what users submit, and tells AI agents what parameters the Actor accepts.

Key-value store

A key-value store is storage for unstructured or auxiliary data, such as Actor input, files, screenshots, and images. Each run's store holds its INPUT record and any files the Actor writes, and its layout can be described with a key-value store schema.

MCP connectors

MCP connectors let an Actor call third-party services, such as Notion, Slack, and GitHub, over the Model Context Protocol using credentials you authorize once in your account settings. They work in the opposite direction to the Apify MCP server, which exposes Actors to outside AI clients.

Output

Output is the data an Actor run produces. Structured results go to the run's dataset and anything else to its key-value store, and you read both from Apify Console or the Apify API.

Output schema

An output schema is the definition of where an Actor stores its results and how Apify Console and the run API endpoint present them. Publishing an Actor requires one.

Pay-per-event

Pay-per-event (PPE) is a pricing model that charges users for specific events an Actor emits, such as each result or action, rather than for compute time. Its pay-per-event plus usage option also charges users the Actor's platform usage costs, which lowers its quality score.

Pay-per-usage

Pay-per-usage is one of the two pricing models on Apify Store. Users pay only the usage costs an Actor generates on the Apify platform.

Proxy

A proxy is a server that forwards your requests, so target websites see its IP address instead of yours. On the Apify platform, you get proxies through Apify Proxy as datacenter, residential, or Google SERP proxies.

Quality score

A quality score is a 0 to 100 rating of an Apify Store Actor, built from categories such as reliability, popularity, ease of use, and pricing transparency. It influences how the Actor ranks in Apify Store.

Request queue

A request queue is storage for the URLs to crawl that tracks their state. It deduplicates URLs and records which are pending or handled, so a crawl can resume after an interruption and be shared across runs.

Residential proxy

A residential proxy routes requests through IP addresses of real home devices, which makes it harder to detect and block than a datacenter proxy. It draws on a wider address pool at a higher price, for targets with strong anti-scraping measures.

Schedule

A schedule is a cron-based trigger that runs an Actor or task at set intervals. You manage it from Apify Console or the Apify API.

Webhook

A webhook is an HTTP callback the Apify platform sends when an Actor run changes state. Use it to trigger an external system, or another Actor, when a run succeeds or fails, instead of polling for the result.