Agentic MCP Client

🚀 Automated MCP Client Agent

An independent proxy runner that uses MCP (Model Context Protocol) tools to execute tasks via Anthropic Claude, AWS Bedrock, and OpenAI APIs. It enables AI agents to run autonomously in a cloud environment and securely interact with various systems.

🚀 Quick Start

Run the Dashboard Web

cd dashboard
npm i
npm run dev

Dashboard URL: http://localhost:3000
API Documentation: http://localhost:3001

Installation Steps

1. Clone the repository:

git clone [Repository Link]
cd agentic-mcp-client

2. Set up dependencies:

npm install

3. Create a configuration file:

touch src/config.json

✨ Features

• Includes a basic agent dashboard.
• Uses JSON configuration files to define tasks for running independent agents.
• Supports Anthropic Claude and OpenAI models.
• Session logging to track agent progress.

💻 Usage Examples

Configuration Guide

Example Configuration

Inference Server

"inference_server": {
  "type": "openai",
  "api_key": "your_openai_api_key",
  "model_name": "gpt-3.5-turbo"
}

MCP Tools

"mcp_tools": [
  {
    "name": "tool1",
    "description": "Execute Task 1",
    "uri": "http://localhost:8080/tool1"
  },
  {
    "name": "tool2",
    "description": "Execute Task 2",
    "uri": "http://localhost:8081/tool2"
  }
]

How MCP Works

sequenceDiagram
    participant User
    participant AgentWorker
    participant LLM as Language Model
    participant MCP as MCP Tools

    User->>AgentWorker: Task + Configuration
    AgentWorker->>MCP: Initialize Tools
    AgentWorker->>LLM: Send Task
    loop Until Completion
        LLM->>AgentWorker: Request Tool Usage
        AgentWorker->>MCP: Execute Tool
        MCP->>AgentWorker: Tool Result
        AgentWorker->>LLM: Send Tool Result
        LLM->>AgentWorker: Response
    end
    AgentWorker->>User: Final Result

🤝 Contribution Guide

We welcome contributions to the Agentic MCP Client! To contribute, follow these steps:

1. Fork the repository.
2. Create a new branch for your feature or bug fix.
3. Make changes and commit them.
4. Push the changes to your forked repository.
5. Create a pull request to the main repository.

🙏 Acknowledgments

This project is inspired by outstanding open-source projects in the MCP ecosystem:

• MCP-Bridge - A middleware that provides an OpenAI endpoint compatible with MCP tools, helping us understand methods for tool integration and standardization.

Thanks to the contributors of these projects for their pioneering work in the MCP field, making agent development more versatile and powerful.

📄 License

The Agentic MCP Client is released under the Apache 2.0 license. For more details, see the LICENSE file.