Report Abuse

Basic Information

MCPAdapt is a developer-focused adapter library that makes Model Context Protocol (MCP) servers available as tools inside popular agentic frameworks. It lets agent developers call and integrate 650+ existing MCP servers and their tools into frameworks like Smolagents, LangChain, CrewAI and Google GenAI without reimplementing connectors. The project provides a core MCPAdapt context manager that can wrap one or multiple MCP servers provided as local stdio commands or remote SSE endpoints and convert their JSON-described tools into framework-native tool objects. The README and examples show how to instantiate server parameters, choose an adapter, and obtain a flattened tool collection to use directly in agent workflows. The repo targets developer workflows for building or extending agents rather than end-user applications.

Links

Categorization

App Details

Features
Adapters for multiple agent frameworks with optional extras to avoid unnecessary dependencies. Built-in support and example adapter for Smolagents and adapters or guides for LangChain, CrewAI and Google GenAI. Ability to adapt one or many MCP servers at once and flatten resulting tool lists. Support for local stdio server parameters and remote MCP servers via Server-Sent Events. Synchronous and asynchronous adapter paths when the underlying framework supports async. Easy installation via pip extras and integration into Smolagents 1.4.1+. Documentation, examples, a roadmap listing planned adapters, and a tests workflow are included.
Use Cases
MCPAdapt reduces integration work for agent developers by turning existing MCP servers into ready-to-use tools inside agent frameworks, so developers can leverage many community-contributed servers without writing custom wrappers. It enables combining multiple MCP servers into a single tool collection, supports both local and remote servers, and provides async handling when frameworks permit it. The adapter pattern lets teams add new framework support with a small adapter module, enabling reuse and consistent tooling across agent projects. Documentation, examples and a security warning about untrusted MCP servers help teams adopt the library safely and quickly.

Please fill the required fields*