Apify Model Context Protocol (MCP) Server

The Apify MCP Server enables interaction with Apify Actors via the Model Context Protocol, allowing AI assistants to use them as tools. It supports two modes: an HTTP server accessible via Server-Sent Events (SSE) and a local server via standard input/output (stdio). Clients like Claude Desktop, LibreChat, and Apify Tester MCP Client can interact with the server.

The server allows AI agents to utilize Apify Actors for tasks like web scraping, data extraction, and more. It includes helper tools for discovering Actors and managing the tool list. The server is pre-configured with Actors like apify/instagram-scraper and lukaskrivka/google-maps-with-contact-details.

Usage involves running the server as an Apify Actor or locally. The Actor runs in Standby mode with an HTTP web server that receives and processes requests. Local setup requires Node.js and an Apify API token. Debugging is facilitated by the MCP Inspector. The roadmap includes adding Apify's dataset and key-value store as resources.