Local MCP server experiments and tooling for macOS.
mq-mcp
Local MCP server experiments and tooling for macOS.
mq-mcp is a small local-first lab for building, testing, and documenting MCP-related workflows on macOS. The goal is to make local MCP setup easier to understand, repeat, validate, and publish safely.
Status
Early prototype with local MCP tools, validation, and repo-aware helpers.
This repository is currently useful as:
- a local MCP server experiment
- a documentation baseline for MCP setup on macOS
- a place to collect repeatable setup, validation, and troubleshooting flows
- a publishable project shell with README, roadmap, changelog, license, release, and GitHub Pages docs
It is not yet a polished package or production-ready MCP distribution.
What is inside
| Path | Purpose |
| --- | --- |
| mq-mcp/server.py | Local FastMCP server experiment |
| mq-mcp/bridge.py | Bridge between OpenAI and the local MCP server |
| mq-mcp/main.py | Minimal Python entry point |
| mq-mcp/pyproject.toml | Python project metadata and dependencies |
| mq-mcp/.env.example | Example environment file |
| docs/index.html | GitHub Pages landing page |
| docs/screenshots/ | Place for screenshots and visual docs |
| CHANGELOG.md | Release history |
| ROADMAP.md | Planned work |
| VERSION | Current project version |
| release.sh | Local release helper |
| docs/install.md | macOS installation guide |
Quick Start
Clone the repository and enter the Python project folder:
git clone https://github.com/MCamner/mq-mcp.git
cd mq-mcp/mq-mcp
Install dependencies with uv:
uv sync
Run the minimal Python entry point:
uv run python main.py
Run the local MCP server experiment:
uv run mcp run server.py
Run the bridge with a prompt:
uv run python bridge.py "List the available MCP tools."
Requirements
The Python project currently declares:
- Python
>=3.14 mcp[cli]openaipsutilpandaspillowpyguitarprorequests
If your local Python version does not match the project requirement, use uv to manage the environment.
Environment
Copy the example environment file before running anything that needs API credentials (run from inside mq-mcp/mq-mcp/):
cp .env.example .env
Do not commit real API keys, tokens, private paths, or secrets.
GitHub Pages
Live docs:
https://mcamner.github.io/mq-mcp/
Current docs include:
- a GitHub Pages landing page
- screenshot folder structure
- public readiness baseline
- macOS installation guide in
docs/install.md
Demo material
Current demo material:
- macOS installation guide in
docs/install.md - GitHub Pages landing page in
docs/index.html - local MCP server experiment in
mq-mcp/server.py - OpenAI/MCP bridge experiment in
mq-mcp/bridge.py
Planned demo material:
- local MCP server startup output
- example MCP configuration
- validation command output
- troubleshooting flow
- screenshots of setup and usage
Safety notes
See docs/security.md for the MCP safety policy.
This project is local-first and experimental.
Before using or extending it:
- review what each MCP tool can access
- avoid committing
.envfiles - keep filesystem access scoped and explicit
- avoid hardcoded machine-specific paths where possible
- prefer read-only tools until validation is solid
- document every command that touches local files or credentials
Available MCP tools
The local MCP server currently exposes these tools:
get_system_resources— shows CPU, memory, and disk informationread_repo_file— reads a file inside the repository rootlist_repo_files— lists repository files up to a chosen depthsearch_repo— searches repository textgit_status— shows branch, working tree status, and recent commitsgit_diff— shows current git diff, optionally for one pathvalidate_project— runsscripts/validate.shwhen availableupdate_repo_file— safely replaces exact text in allowed repo files without committingrun_mqlaunch— runsmqlaunch.shanalyze_csv— analyzes CSV filesanalyze_guitar_pro— analyzes Guitar Pro filesopen_in_app— opens a repo file in the default appedit_image— edits an image with supported actions
Validation
Run the local validation script from the repository root:
./scripts/validate.sh
The validation checks:
- required project files
- absence of debug and backup files
- Python syntax compilation
- MCP tool listing
- core MCP tools including
read_repo_file,list_repo_files,search_repo,git_status,git_diff,validate_project, andupdate_repo_file
You can also run validation through the bridge:
cd mq-mcp
uv run python bridge.py "Run project validation."
Development checks
Useful local checks:
git status
python -m compileall mq-mcp/
Run the safety tests:
uv --directory mq-mcp run pytest tests/ -v
Roadmap
See ROADMAP.md for planned MCP setup, validation, troubleshooting, documentation, and release work.
License
See LICENSE.