MCP Ectors

🚀 Enterprise MCP SSE Server (mcp-ectors) Introduction

The Enterprise's MCP SSE Server (mcp-ectors) is a cost - effective, scalable, and user - friendly platform designed specifically for enterprises. This system aims to simplify the integration of large language models (LLMs) with tools and resources through actors, the Rust language, and a concise multi - router management architecture, facilitating advanced agent - based AI workflows.

🚀 Quick Start

✨ Features

• High - performance and scalable architecture for enterprise - level applications.
• Simplify the integration of LLMs with tools and resources.
• Support for multiple communication protocols and dynamic function expansion.

📦 Installation

No installation steps are provided in the original document, so this section is skipped.

💻 Usage Examples

No specific code usage examples are provided in the original document, so this section is skipped.

📚 Documentation

🔧 Technical Details

Architecture Overview

1. Server Builder

   • server_builder is responsible for determining the transport layer. Currently, it supports SSE (Server - Sent Events), and future versions will add stdio and wasi transports.
   • The current version only supports SSE as the communication medium between the client and the server.

2. Router Service Manager

   • RouterServiceManager is responsible for registering multiple routers and ensuring that each router can handle requests independently without establishing new connections.
   • This architecture allows multiple routers to be deployed under the same connection, significantly improving the system's efficiency and scalability.

3. Log Configuration

   • The server can be configured to store logs in a specified directory and supports custom log levels (such as INFO, DEBUG, etc.) for monitoring and debugging.
   • Example configuration:

     let log_config = LogConfig {
         log_dir: "logs".to_string(),
         log_file: "server.log".to_string(),
         level: Level::INFO,
     };
     

4. Transport Role

   • Any transport used for the MCP server needs to implement the TransportActorTrait. This trait ensures that different transport methods (such as SSE, stdio, wasi) can correctly handle requests and messages.
   • Key message types:
     • TransportRequest: Allows new JsonRPCMessages to be sent to the client and is responsible for communication between the server and the client.
     • StartTransport: Used to start a specific transport layer and initialize related background tasks (such as establishing a connection or waiting for inbound messages).
     • StopTransport: Used to stop the transport layer and gracefully close any active connections or associated tasks.

5. Router

   • Router is a feature interface that all routers need to implement and is closely designed around the MCP standard. After registering a new route through RouterRegistry, each route is encapsulated in a specific RouterActor responsible for handling corresponding functional requests.
   • For example, when a tools/call request is received, the RouterActor holding the corresponding router will respond according to the preset logic. This dynamic response mechanism enables the server to flexibly support various functional expansions.

6. Standard Actors

   • MCP defines a series of standard actors responsible for implementing basic functions such as initialization and management of tools, prompts, and resources.
   • These standard actors handle initial requests and response list requests (such as listing tools, prompts, resources), making interactions with these core components unified and standardized.

Future Development Plans

• MCP Protocol Extension The basic functions have been implemented, but the notification mechanism is still in planning. In addition, future versions will also support functions such as authentication/authorization, error handling, and logging.

• Transport Extension Plans to add support for stdio and wasi to enrich the server's communication methods.

• Seeking Help The community is welcome to contribute code or provide feedback. If you find any issues, please submit them through the following link: GitHub
  Contact: Zhang San

How to Contribute

We warmly welcome the active participation of the open - source community! You can contribute in the following ways:

1. Report Issues: Submit an Issue in the GitHub repository.
2. Feature Development: Fork the repository and implement new features in a new branch.
3. Code Review: Participate in the review of existing Pull Requests.

📄 License

No license information is provided in the original document, so this section is skipped.

Contact Us

• Email: Enterprise Support Team
• Phone: +86 - 010 - 12345678

Join us to drive the development of AI technology together!