apify/apify-mcp-server

What it solves

Apify MCP Server allows AI agents to access thousands of ready-made web scrapers, crawlers, and automation tools (Actors) from the Apify Store. It bridges the gap between LLMs and the live web, enabling agents to extract structured data from social media, search engines, maps, and e-commerce sites without the developer having to manually build each integration.

How it works

The project implements the Model Context Protocol (MCP), allowing it to connect to compatible clients like Claude Desktop, Cursor, and VS Code. It can be deployed as a hosted HTTPS endpoint or run locally via standard input/output (stdio).

Key mechanisms include:

• Dynamic Tool Discovery: Agents can search for and retrieve the input schemas of any Apify Actor, allowing the LLM to understand how to call the tool and what arguments to provide.
• Actor Execution: The server translates MCP tool calls into Apify Actor runs, returning metadata and storage IDs that the agent can then use to fetch final results.
• Agentic Payments: It supports machine-to-machine payment methods (AGI, x402, and Skyfire) so autonomous agents can pay for Actor runs without requiring a traditional user account.

Who it’s for

Developers and AI researchers building AI agents that need real-time web data extraction, automation, and the ability to dynamically discover and use new web-scraping tools.

Highlights

• Massive Tool Library: Provides instant access to 8,000+ Apify Actors for diverse data extraction tasks.
• Dynamic Schema Inference: Automatically loads Actor input schemas so LLMs know exactly how to interact with them.
• Flexible Deployment: Supports both a hosted server (via OAuth) and local stdio execution.
• Autonomous Payments: Integrated support for x402 and Skyfire for agent-led financial transactions.