The actors-mcp-server is an MCP server designed to seamlessly connect AI models with Apify Actors, enabling them to perform specific tasks. It acts as a bridge, allowing AI assistants to leverage Apify's vast library of Actors for web scraping, data extraction, and more.

This server supports both HTTP/SSE and standard input/output (stdio) for versatile integration with various MCP clients like Claude Desktop and LibreChat. By exposing Apify Actors as tools, it empowers AI agents to access real-world data and functionalities, enhancing their problem-solving capabilities.

Key benefits include simplified access to Apify's ecosystem, automated data handling, and the ability to build sophisticated AI workflows. Developers can easily integrate this server into their AI applications, unlocking a new realm of possibilities for AI-driven automation and data analysis. It's implemented as an Apify Actor, configurable via the Apify platform or locally, offering flexibility and ease of deployment.

Actor Integration as Tools

The actors-mcp-server allows AI models to leverage any Apify Actor as a tool, effectively extending the model's capabilities to perform specific tasks. This is achieved by dynamically loading Actor input schemas and creating corresponding MCP tools. The tool name is the full Actor name (e.g., apify/rag-web-browser), and the arguments for the tool represent the input parameters of the Actor. This integration enables AI models to interact with the real world through Apify's vast library of web scraping and automation tools. For example, an AI model can use the apify/facebook-posts-scraper Actor to extract data from Facebook posts or the lukaskrivka/google-maps-with-contact-details Actor to extract contact details from Google Maps. The server manages the execution of these Actors, passing the arguments provided by the AI model as input parameters.

Dynamic Actor Discovery

The server provides helper tools to discover available Actors and retrieve their details, enhancing the AI model's ability to adapt and utilize new tools. The get-actor-details tool retrieves documentation, input schema, and details about a specific Actor, while the discover-actors tool searches for relevant Actors using keywords. This dynamic discovery feature allows AI models to explore and integrate new functionalities without requiring manual updates or redeployment. For instance, an AI model can use the discover-actors tool to find Actors related to "e-commerce data extraction" and then use get-actor-details to understand how to use the most relevant Actor. This capability is particularly useful in rapidly evolving environments where new tools and data sources are constantly emerging.

Standard Input/Output (stdio) Support

The actors-mcp-server supports interaction via standard input/output (stdio), enabling integration with clients like Claude Desktop that may not support Server-Sent Events (SSE). This allows developers to run the MCP server locally and connect it to AI models for testing and development purposes. The server provides example clients (clientStdio.ts and clientStdioChat.ts) that demonstrate how to connect to the MCP server, list available tools, and call specific tools using stdio transport. For example, a developer can use the clientStdioChat.ts script to interact with the server in a conversational manner, testing different tool calls and responses. This feature is crucial for developers who prefer a local development environment or need to integrate with clients that have limited transport support.