Booner_MCP

An AI infrastructure-as-code platform using Model Context Protocol (MCP) with Ollama for agentic coding and server management.

Overview

This project allows AI agents to interact with local infrastructure, deploy and manage various server types (web servers, game servers, databases) through the Model Context Protocol. It integrates with a local Ollama deployment running mixtral on powerful hardware.

System Architecture

Machines Configuration

• Machine 1 (Booner_Ollama - 10.0.0.10):

  • Runs Ollama LLM server on port 11434
  • Provides AI capabilities to the MCP system
  • Hardware: AMD 5700X3D, 4070 Ti Super, 64GB RAM, Quadro P4000

• Machine 2 (Booner_MCP & agent - 10.0.0.4):

  • Runs the MCP core management API
  • Runs the Next.js web interface
  • Hardware: Ryzen 7 5700X3D, 32GB RAM
  • Worker Agent for Agentic Coding

• Machine 3 (booner-mcp-web & agent - 10.0.0.4):

  • Runs the Web GUI to interact with MCP & Ollama
  • Worker Agent for Agentic Coding

• Machine 3 (OPN_IaC - 10.0.0.2):

  • Runs infrastructure as code tools
  • Dedicated OPNSense Network API

• Deployment Targets (Machine N):

  • Run game servers, web applications, etc.
  • Hardware examples:
  • E5-2680 v4, 16GB RAM, RTX 3050 8GB

• Storage: TrueNAS with 8TB HDD & 2TB SSD (NFS shared)

Software Stack

• OS: Ubuntu 24 (all machines)
• LLM: Mixtral via Ollama
• Primary Languages: Python, Go, NextJS
• Containerization: Docker

Project Structure

• agents/: AI agent definitions and orchestration code
• servers/: MCP server implementations for different infrastructure tasks
• api/: API server for agent communication
• config/: Configuration files for different environments and systems

Setup & Deployment

Prerequisites

• Git with support for submodules
• Docker and Docker Compose
• Node.js 18+ (for local development)
• Python 3.11+ (for local development)

Initial Setup

1. Clone the repository with submodules:

   git clone --recurse-submodules https://github.com/vespo92/Booner_MCP.git
   cd Booner_MCP
   

2. Create an environment file:

   cp .env.example .env
   

3. Generate a secure AUTH_SECRET:

   # On Linux/macOS
   ./generate_auth_secret.sh
   
   # On Windows
   .\generate_auth_secret.ps1
   

4. Deploy with Docker Compose:

   docker-compose up -d
   

Accessing the Services

• Web UI: http://10.0.0.1:3000
• API: http://10.0.0.1:8000
• Ollama: http://10.0.0.10:11434

Development Workflow

Working with the Main Project

cd Booner_MCP
# Make changes
git add .
git commit -m "Your commit message"
git push origin main