mcpField
M
MCP Servers

MCP Servers

Local
github.com/ajoslin103Developer Tools0Updated 84w ago
Repository

About MCP Servers

MCP Servers configures a collection of MCP servers with Docker Compose and includes scripts for their lifecycle and a local MCP Inspector. It is intended for setting up and managing Git, filesystem, PostgreSQL, and memory services together.

Key features

  1. 01Docker Compose server setup
  2. 02Git repository operations
  3. 03Controlled filesystem access
  4. 04PostgreSQL schema inspection
  5. 05Persistent knowledge graph memory
  6. 06Local MCP Inspector

Use cases

  1. 01Run several MCP servers together
  2. 02Inspect database schemas through PostgreSQL
  3. 03Manage and debug servers with Inspector
  4. 04Add a local server using script templates

MCP support

It runs Git, Filesystem, PostgreSQL, and Memory MCP servers through Docker Compose, with MCP Inspector for local management and debugging. Script templates support adding locally run servers.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
  "mcpServers": {
    "servers-ajoslin103": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "compose"
      ]
    }
  }
}
Run commanddocker run -i --rm compose

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
ocicompose

Tools

No tool list yet

Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.

Details

From the project README

MCP Servers

This project sets up Model Context Protocol (MCP) servers using Docker Compose.

Included Servers

  • Git - Read, search, and manipulate Git repositories
  • Filesystem - Secure file operations with configurable access controls
  • PostgreSQL - Database access with schema inspection
  • Memory - Knowledge graph-based persistent memory system
  • Inspector - GUI tool for managing and debugging MCP servers (run locally with npx)

Setup

  1. Initialize the repository:
./scripts/init-repo.sh

This will:

  • Set the correct Node.js version using nvm
  • Install MCP Inspector globally
  1. Start the containerized servers:
docker compose up -d
  1. Start the MCP Inspector locally:
./scripts/start-inspector.sh
  1. Stop the MCP Inspector:
./scripts/stop-inspector.sh
  1. Stop the containerized servers:
docker compose down

Directory Structure

  • docker-data/ - Data directories for each MCP server
  • docker-files/ - Dockerfile resources
  • scripts/ - Utility scripts for managing services
    • init-repo.sh - Initialize the repository with correct Node.js and packages
    • start-inspector.sh - Start the MCP Inspector in background
    • stop-inspector.sh - Stop the MCP Inspector
    • templates:
    • template-start-mcp-server.sh - Template for creating start scripts for MCP servers
    • template-stop-mcp-server.sh - Template for creating stop scripts for MCP servers

Configuration

Each server can be configured through environment variables in the docker-compose.yml file.

Accessing the Inspector

The MCP Inspector will be available at http://localhost:5173 after starting it with the script.

Logs

  • Inspector logs are stored in ./docker-data/inspector/inspector.log

Adding New MCP Servers

To add a new MCP server that runs locally (non-dockerized):

  1. Copy the template scripts:
cp scripts/template-start-mcp-server.sh scripts/start-newserver.sh
cp scripts/template-stop-mcp-server.sh scripts/stop-newserver.sh
chmod +x scripts/start-newserver.sh scripts/stop-newserver.sh
  1. Edit both scripts to set:
  • SERVER_NAME - Name of your server
  • NPX_COMMAND - NPX command to run the server
  • PORT - Port the server runs on
  1. Make sure to install the server package globally or specify it in the NPX command.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.