MCPAdapt unlocks access to 650+ Model Context Protocol (MCP) servers for agentic frameworks, enabling seamless integration of diverse data and tools into workflows. It supports frameworks like Langchain, CrewAI, and Smolagents.
MCPAdapt bridges the gap between Model Context Protocol (MCP) servers and agentic frameworks, granting access to 650+ tools. It adapts MCP servers into usable tools within frameworks like Smolagents, Langchain, and CrewAI.
Installation involves specifying the desired framework during package installation (e.g., uv add mcpadapt[langchain]). Smolagents 1.4.1+ has native integration.
Usage involves instantiating MCPAdapt with server parameters (command or SSE details) and the desired adapter. The adapter converts MCP tools into the framework's native tool format.
Security Warning: Exercise caution with MCP servers, especially over SSE, due to potential remote code execution risks.
The project encourages contributions by creating new adapters for unsupported frameworks. The roadmap includes support for pydantic-ai, llamaindex, and swarm. It already supports remote MCP Servers via SSE, jupyter notebook and has tests.