mcp-protocol-sdk
[DEPRECATED] Moved to prism-mcp-rs - Enterprise-grade Rust MCP SDK
claude mcp add --transport stdio prismworks-ai-mcp-protocol-sdk docker run -i mcp-rust/mcp-protocol-sdk:latest
How to use
The MCP Protocol SDK is a production-ready Rust implementation of the Model Context Protocol. It provides tools to build MCP servers, define tools and resources, and expose them over multiple transports (including STDIO, HTTP, and WebSocket). The repository emphasizes a complete development flow with tests, documentation, and examples, making it suitable for embedding AI-driven capabilities into external systems. Typical use involves creating an MCP server instance, registering tools with their schemas and handlers, and launching the server over a transport compatible with your environment (for example STDIO for Claude Desktop or HTTP/WebSocket for remote clients). The examples in the repository show a Calculator-based tool handler and demonstrate how to construct ToolHandler implementations, expose them via the MCP server, and start the server using the provided transports. You can also build clients that connect to the server, query capabilities, and invoke tools, all while benefiting from the SDK’s type-safe, async design and built-in validation and error handling.
How to install
Prerequisites:
- Rust toolchain (rustup, cargo) installed on your system
- Basic familiarity with Rust and Cargo projects
Installation steps:
- Install Rust and cargo (if not already installed):
- On Windows: download from https://www.rust-lang.org/tools/install
- On macOS: brew install rustup; rustup
- On Linux: curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
- Clone the repository or use crates.io (this SDK is published as a crate):
- git clone https://github.com/mcp-rust/mcp-protocol-sdk.git
- cd mcp-protocol-sdk
- Build from source (recommended for development):
- cargo build --release
- Run tests (optional but recommended):
- cargo test
- If you want to use the crate in a project, add it as a dependency in your Cargo.toml: [dependencies] mcp-protocol-sdk = "0.5.0"
- For quick experimentation, you can also explore examples in the repo (e.g., building a server with a tool handler) and adapt the code into your own application.
Additional notes
Notes and tips:
- The SDK is designed to be async with Tokio; enable appropriate features for your transports (e.g., stdio, http) when building your project.
- It ships with extensive documentation, examples, and tests to help you implement servers, clients, and transports that conform to the MCP specification.
- The MCP schema is 100% compliant with the 2025-06-18 MCP specification in this SDK; ensure your client and server versions align with this spec.
- If you plan to expose a server over HTTP/WebSocket, consider performance tuning (connection timeouts, max message size) and enabling TLS where appropriate.
- When using the docker-based run approach, make sure the image contains the necessary runtime and compiled binary compatible with your host architecture.
- Common environment variables you might configure include logging level (RUST_LOG), server name/version, and transport-specific options. Add placeholders like RUST_LOG=info or replace with your desired configuration.
Related MCP Servers
penpot
Penpot MCP server
neurolink
Universal AI Development Platform with MCP server integration, multi-provider support, and professional CLI. Build, test, and deploy AI applications with multiple ai providers.
offeryn
Build tools for LLMs in Rust using Model Context Protocol
notebooklm -secure
Secure NotebookLM MCP Server - Query Google NotebookLM from Claude/AI agents with 14 security hardening layers
goai
AI SDK for building AI-powered applications in Go
commune
A Rust library designed to support the development of discoverable networks of AI agents.