Apify MCP Server

Run thousands of ready-made Apify Actors to scrape social media, search engines, maps, and ecommerce sites, plus access docs and stored datasets.

Data & Enrichment by Apify OAuth2, Bearer Token, API Key active

Connect (Remote Server) GitHub Docs npm: @apify/actors-mcp-server

Overview

The Apify MCP Server gives AI agents access to the Apify Store, a marketplace of thousands of pre-built web scrapers, crawlers, and automation tools called Actors. Through a single MCP connection, an agent can discover relevant Actors, inspect their input schemas, execute them, and retrieve the resulting data from Apify datasets or key-value stores. Targets include social platforms (Instagram, TikTok, X, Facebook), search engines (Google, Bing), maps, ecommerce listings (Amazon, eBay, Shopify), and generic websites via the RAG Web Browser Actor.

The server is offered as a hosted endpoint at https://mcp.apify.com with OAuth or Bearer token auth, and as a local stdio server published on npm as @apify/actors-mcp-server. By default it exposes actor discovery and execution tools, Apify documentation search, and the RAG Web Browser. Additional categories like runs, storage, and preview can be enabled via the --tools CLI flag or the ?tools= query parameter. The server also supports dynamic tool discovery for clients that can register new tools at runtime, exposing add-actor instead of call-actor.

Notably, the server supports agentic payments through x402 (USDC on Base) and Skyfire (PAY tokens), allowing agents to autonomously pay for paid Actors. It is officially maintained by Apify, MIT licensed, and works with Claude Desktop, Claude.ai, Cursor, VS Code, ChatGPT, and any other MCP-compatible client.

Tools

Tool	Description
`search-actors`	Search the Apify Store for Actors by keyword, category, or use case.
`fetch-actor-details`	Get a specific Actor's description, input schema, pricing, and usage information.
`call-actor`	Execute an Apify Actor with provided input and return its results.
`add-actor`	Dynamically register an Actor as a new tool (for clients that support runtime tool discovery).
`get-actor-output`	Retrieve the output dataset of a completed Actor run.
`search-apify-docs`	Search the Apify documentation for relevant pages.
`fetch-apify-docs`	Fetch the full content of an Apify documentation page.
`apify-slash-rag-web-browser`	Browse the web and return clean, LLM-ready content via the Apify RAG Web Browser Actor.
`get-actor-run`	Get detailed information about a specific Actor run. Requires the runs tool category.
`get-actor-log`	Retrieve logs from an Actor run. Requires the runs tool category.
`get-dataset-items`	Retrieve items from an Apify dataset. Requires the storage tool category.
`get-key-value-store-record`	Get a value from an Apify key-value store by key. Requires the storage tool category.

Setup Guide

Prerequisites

An Apify account and API token from console.apify.com/account/integrations
Node.js v20+ if running locally
An MCP-compatible client (Claude Desktop, Cursor, VS Code, etc.)

Option 1: Hosted server (recommended)

Point your MCP client at https://mcp.apify.com. The endpoint supports OAuth or a Bearer token.

{
  "mcpServers": {
    "apify": {
      "url": "https://mcp.apify.com"
    }
  }
}

To enable additional tool categories, pass them via the tools query parameter:

https://mcp.apify.com?tools=actors,docs,runs,storage

Option 2: Local stdio server

Install and run via npx with your Apify token in the environment:

{
  "mcpServers": {
    "apify": {
      "command": "npx",
      "args": ["-y", "@apify/actors-mcp-server"],
      "env": {
        "APIFY_TOKEN": "your-apify-token-here"
      }
    }
  }
}

Enable extra tool categories with the --tools flag:

npx @apify/actors-mcp-server --tools actors,docs,runs,storage

Default tools enabled

Without configuration, the server exposes the actors category (search-actors, fetch-actor-details, call-actor), the docs category (search-apify-docs, fetch-apify-docs), and the apify/rag-web-browser Actor. Categories like runs, storage, and preview must be enabled explicitly.

Use Cases

Scrape Instagram, TikTok, or X profiles and posts for competitive research, then summarize trends in a report
Pull Google Maps listings (reviews, ratings, contact info) for a list of cities to build a sales prospect database
Extract product data from Amazon, Shopify, or eBay listings for price monitoring and market analysis
Crawl a website and feed clean, LLM-ready content to an agent via the RAG Web Browser Actor
Discover the right Apify Actor on demand ("find me a LinkedIn jobs scraper") and run it with structured inputs

Example Prompts

"Find an Apify Actor that scrapes Google Maps reviews, then pull all reviews for the 10 highest-rated coffee shops in Brooklyn."
"Search Apify for a TikTok scraper and get the latest 50 videos from @nasa."
"Use the RAG Web Browser to fetch https://example.com/pricing and extract every plan and price."
"Look up the input schema for apify/instagram-scraper and call it to collect posts from three competitor accounts."
"Get the dataset items from my last Actor run and summarize the top trending products."

Pros

Officially maintained by Apify, with both a hosted OAuth endpoint and an open-source npm package
Massive tool surface: any of the thousands of Actors in the Apify Store becomes accessible to the agent
Dynamic discovery means agents can find new scrapers at runtime instead of being limited to a fixed tool set
Supports agentic payments via x402 and Skyfire for autonomous use of paid Actors

Limitations

Requires an Apify account, and most non-trivial scraping consumes paid compute units
Many advanced tools (runs management, storage access) are off by default and need explicit --tools configuration
Actor execution can be slow compared to typical API calls, since scrapers spin up real browsers or workers

Alternatives

Firecrawl MCP Server for general-purpose web scraping and crawling
Bright Data MCP Server for proxy-backed scraping at scale
Browserbase MCP for headless browser automation when you need full control over the page