Skip to content

assagman/mcp-servers

Repository files navigation

mcp-servers

A Python package providing a collection of Model-Control-Protocol (MCP) servers and a CLI tool to manage them efficiently.

Disclaimer

This project is created for personal use and does not guarantee stable behavior. It is made public solely as a reference for other programmers. The project is currently in early development, potentially unstable, and may produce undesired outcomes. Use at your own risk.


Overview

mcp-servers implements Model-Control-Protocol servers for various integrations:

  • Filesystem access
  • Brave Search integration
  • SearXNG search integration
  • Tavily Search integration

These servers can be used by AI agents to interact with your system and external services in a controlled manner.

Pre-requisites

  • Python 3.12+
  • uv (optional but recommended package/environment manager)
  • podman or docker for container operations (e.g., running local SearXNG instance)
  • OpenRouter API key and credits (for experimentation with examples)

Installation

Package Managers

Using uv (recommended)

uv venv --python 3.12
source .venv/bin/activate
uv pip install --upgrade mcp-servers

Using pip

pip install --upgrade mcp-servers

Development Setup

git clone git@github.com:assagman/mcp_servers.git
cd mcp_servers
uv venv --python 3.12
source .venv/bin/activate
uv sync --extra dev

Configuration

The package requires specific configuration files in your home directory. Initialize everything at once with:

mcpserver init

This command will:

  • Create ~/.mcp_servers/.env → MCP Server HOST:PORTs, API KEYs, URLs, etc.
  • Create ~/.mcp_servers/searxng_config/settings.yml → Configuration for local SearXNG instance

⚠️ Important: You must set your own API keys in the generated .env file. They are left empty by default.

For environment variable reference, see .env.example

Usage

CLI Tool

This package provides a CLI tool (mcpserver) to manage configuration, MCP servers, and external container operations:

mcpserver -h  # Show help

Server Management

Each MCP server can be started in standard mode or detached mode. Detached mode runs the server in the background.

Filesystem Server
# Start with temporary directory
mcpserver start --server filesystem

# Operate on specific directory
mcpserver start --server filesystem --allowed-dir $(pwd)

# Custom port
mcpserver start --server filesystem --port 8765 --allowed-dir $(pwd)

# Detached mode
mcpserver start --server filesystem --detached
mcpserver stop --server filesystem  # Stop detached server
Brave Search Server

Requires BRAVE_API_KEY environment variable.

# Start server
mcpserver start --server brave

# Custom port
mcpserver start --server brave --port 8766

# Detached mode
mcpserver start --server brave --detached
mcpserver stop --server brave  # Stop detached server
SearXNG Search Server

Requires SEARXNG_BASE_URL environment variable.

# Start local SearXNG container
mcpserver run_external_container --container searxng

# Start server
mcpserver start --server searxng

# Custom port
mcpserver start --server searxng --port 8767

# Detached mode
mcpserver start --server searxng --detached
mcpserver stop --server searxng  # Stop detached server

# Stop SearXNG container
mcpserver stop_external_container --container searxng
Tavily Search Server

Requires TAVILY_API_KEY environment variable.

# Start server
mcpserver start --server tavily

# Custom port
mcpserver start --server tavily --port 8768

# Detached mode
mcpserver start --server tavily --detached
mcpserver stop --server tavily  # Stop detached server

Python API

The package can also be imported and used programmatically. See example files in examples/package_usage

Examples

The package includes examples demonstrating Agent-MCP Server usage with pydantic_ai Agents. All examples use MCPServerHTTP to connect agents with MCP Servers.

To experiment with all MCP servers:

  1. Set BRAVE_API_KEY in ~/.mcp_servers/.env (for Brave search server)
  2. Set OPENROUTER_API_KEY in ~/.mcp_servers/.env (required for all examples)
  3. Start SearXNG container (for SearXNG search server)

CLI Usage Examples

See examples/cli_usage for examples requiring MCP servers to be started via the CLI commands mentioned above.

Package Usage Examples

See examples/package_usage for examples that can be executed as-is.

Advanced Usage

Custom Configuration

You can customize server behavior by modifying configuration files:

# Edit SearXNG settings
vim ~/.mcp_servers/searxng_config/settings.yml

# Edit environment variables
vim ~/.mcp_servers/.env

Multiple Servers

You can run multiple MCP servers simultaneously by specifying different ports:

mcpserver start --server filesystem --port 8765 --detached
mcpserver start --server brave --port 8766 --detached
mcpserver start --server searxng --port 8767 --detached
mcpserver start --server tavily --port 8768 --detached

Troubleshooting

Common Issues

  1. API Keys Not Working: Ensure you've set the correct API keys in ~/.mcp_servers/.env
  2. Port Conflicts: If a port is already in use, specify a different port with --port
  3. Container Issues: Use podman logs searxng or docker logs searxng to diagnose SearXNG container problems

Testing Specifications

  • Tested on macOS (arm64)
  • Python 3.12

License

This project is provided as-is with no warranty. See the LICENSE file for details.

Contributing

Contributions are welcome but not expected. If you find a bug or have a feature request, please consider forking this repo and use your own custom version.

About

A collection of MCP servers and a CLI tool

Topics

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages