Mo5 MCP server

MO5 MCP Server (RAG Connector)

This MCP server (Model Context Protocol) acts as a bridge between modern AI agents (Claude Desktop, Augment, Cursor) and the Thomson MO5 ecosystem. It allows access to an expert knowledge base and manipulation of MO5 heritage files.

It provides semantic search tools, documentation resources, and expert prompts to facilitate development around the Thomson MO5 microcomputer.

🏗️ MO5 Development Ecosystem

This server is part of a complete toolchain for modern 6809 development:

C library optimized for CMOC.

• SDK MO5 : https://github.com/thlg057/sdk_mo5

Project template with automated Makefile.

• MO5 Project Template : https://github.com/thlg057/mo5_template

• This MCP Server : The assistant that connects everything, answers technical questions, and generates assets.

🌐 RAG Infrastructure (Artificial Intelligence)

The server relies on a Retrieval-Augmented Generation architecture to provide accurate answers based on real technical documentation.

🚀 Features

• Semantic Search: Intelligent search within the MO5 technical documentation.
• Resources Explorer: Direct access to the list of documents, tags, and indexing status.
• Expert Prompt: A preconfigured "Expert Thomson MO5" mode for the AI.
• Build Tools: Generate bootable .fd disk images, convert .fd to .sd for SDDrive, and convert PNG assets to C sprite headers.
• Robustness: Native handling of timeouts (30s) and retry policy for unstable network environments (NAS).

🌐 Public MO5 RAG Server

The MO5 RAG server is now publicly deployed and accessible on the Internet.

👉 https://retrocomputing-ai.cloud/

This MCP server is designed to work out of the box with the public RAG instance.

For most users, this means:

• no RAG server to install
• no Docker setup
• no local infrastructure to maintain

Simply configure the MCP server to point to the public URL and start using it from your coding agent.

Self-hosting a RAG server is still possible if you want to:

• experiment with the internals
• customize the knowledge base
• run everything offline

But it is no longer required for normal usage.

🛠️ Requirements

• Node.js (v18 or higher)
• Python 3 with Pillow (pip install Pillow) — required only for the pngtomo5_sprite tool.

The official public MO5 RAG server is available at: https://retrocomputing-ai.cloud/

You can use this public instance directly, there is no need to host your own RAG server.

📦 Installation & Setup

Option 1 — Via npx (recommended, no installation required)

Configure your agent to run the server directly from npm. No git clone or npm install needed.

{
  "mcpServers": {
    "mo5-server": {
      "command": "npx",
      "args": [
        "-y",
        "@thlg057/mo5-rag-mcp"
      ],
      "env": {
        "RAG_BASE_URL": "https://retrocomputing-ai.cloud"
      }
    }
  }
}

Option 2 — From source

git clone https://github.com/thlg057/mo5-mcp-server.git
cd mo5-mcp-server
npm install

Then configure your agent:

{
  "mcpServers": {
    "mo5-rag": {
      "command": "node",
      "args": ["C:\\your\\path\\to\\mo5-mcp-server\\index.js"],
      "env": {
        "RAG_BASE_URL": "https://retrocomputing-ai.cloud"
      }
    }
  }
}

Note for Windows: Use double backslashes \\ or single slashes / in the path to avoid JSON formatting errors.

🤖 Integration into an Agent (e.g., Augment, Claude Desktop)

Add one of the configurations above to your MCP configuration file (e.g., augmentconfig.json or claudedesktop_config.json).

🧪 Local Testing

It is recommended to test the server before integration to ensure proper communication with the RAG API.

Via the MCP Inspector (Recommended)

# Windows (PowerShell)
$env:RAG_BASE_URL="https://retrocomputing-ai.cloud"; npx @modelcontextprotocol/inspector node index.js

Shortened. The full README is on GitHub.