Get the FREE Ultimate OpenClaw Setup Guide →

neva

Model Context Protocol SDK exploring structured agent infrastructure with explicit runtime behaviour.

Installation
Run this command in your terminal to add the MCP server to Claude Code.
Run in terminal:
Command
claude mcp add --transport stdio romanemreis-neva docker run -i romanemreis/neva:latest

How to use

Neva is a Rust-based MCP server and client SDK designed to be fast and ergonomically configurable. It provides a unified way to build both MCP clients and servers with async, Tokio-powered operation, and supports multiple transports such as stdio for local integrations and Streamable HTTP for remote bidirectional communication. The server side enables you to expose tools, resources, and prompts that can be consumed by MCP clients, while the client side offers convenient APIs to list tools, call tools, and manage resources. Typical usage involves spinning up the Neva server in Rust and interacting with it via the client APIs to define and run your MCP entities (tools, resources, and prompts) in a secure, structured manner. The README highlights example code for creating a sample server with a few simple endpoint-like tools, including a hello tool, a resource getter, and a prompt-based analyzer, all designed to illustrate how to register and expose functionality through the MCP protocol.

To use Neva effectively, you would integrate the Neva library into your Rust project, enable the desired features (e.g., server-full for server capabilities), and run the App with proper stdio or HTTP transports configured. Client usage demonstrates connecting to the server, listing available tools, and invoking tools with arguments. This enables rapid experimentation and iteration when building MCP-based services or integrations.

How to install

Prerequisites:

  • Rust toolchain (rustup, cargo)
  • Git
  • Basic familiarity with Rust projects

Installation steps:

  1. Install Rust (if not already installed): curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh source $HOME/.cargo/env

  2. Clone the Neva repository: git clone https://github.com/RomanEmreis/neva.git cd neva

  3. Build the project (release build recommended for servers): cargo build --release

  4. Run the server binary (the exact binary name may vary; common approach is to run the built executable):

    If a binary named 'neva' or 'neva-server' is produced

    ./target/release/neva-server

    Or, if using a workspace example, you might run:

    cargo run --bin neva-server

  5. Verify installation by checking the server output and ensure transports (stdio or HTTP) are configured as per your environment.

Additional notes

Notes and tips:

  • Neva is designed to stay in sync with the latest MCP specification, so expect potential breaking changes during preview releases.
  • When running locally, prefer the release build for performance.
  • If you plan to expose multiple tools, consider structuring them clearly with appropriate descriptors, URIs, and MIME types for resources.
  • The README demonstrates using stdio transport for local development; you can swap to Streamable HTTP for remote interactions by configuring the server transport accordingly.
  • If you encounter issues with tool discovery or resource resolution, ensure your tool/resource/prompts definitions follow the MCP annotations and descriptions as shown in the examples.
  • Environment variables and configuration can be extended to customize server name, version, and transport endpoints; consult your integration requirements to populate these values.

Related MCP Servers

Sponsor this space

Reach thousands of developers