MCPbundler
MCP server from eugenepyvovarov/MCPbundler
claude mcp add --transport stdio eugenepyvovarov-mcpbundler python -m mcp.server \ --env MCP_LOG_LEVEL="info (set to debug for verbose output)"
How to use
MCP Bundler is a macOS SwiftUI application designed to manage and run MCP servers and skills from a single, native UI. It provides a headless stdio mode for automation, enabling scripted control of MCP lifecycles in addition to the graphical interface. Use the app to add, configure, and start MCP servers, browse and install skills from the built-in marketplaces, and wire up remote endpoints with OAuth flow support. The headless mode allows you to run servers and emit logs to stdout for integration into CI pipelines or automation workflows, making it suitable for local development and automated toolchains.
How to install
Prerequisites:
- macOS (latest stable)
- Xcode (latest stable) for building the GUI if you intend to compile from source
Install / Setup steps:
- Download the MCP Bundler macOS app from the official distribution page or build from source if available.
- Open the app and grant required permissions (network access, file access as prompted).
- In the app, navigate to the Servers tab to add a new MCP server configuration. Provide the command, arguments, and any environment variables as needed for your MCP server.
- If you need headless/CI usage, use the provided headless stdio mode from the app or command-line interface (CLI) options described in the documentation. Configure the server and start it to begin processing MCP workflows.
- To run via the CLI, reference the server configuration you created and invoke the headless runner as needed. Ensure dependencies for your MCP server are installed in the environment that runs the CLI.
Note: If you are building from source, ensure you have Swift and the required Xcode toolchain installed, and follow the repository’s build instructions to compile the macOS app.
Additional notes
Tips:
- The app supports both GUI management and a headless stdio runner for automation; use the headless mode in scripts to integrate with your existing tooling.
- If you enable OAuth or remote authentication features, ensure your callback URLs and scopes are correctly configured in the remote provider settings.
- Logs and health status are available in the UI; enable verbose logging in the environment to aid debugging during setup.
- When configuring env vars for MCP servers, keep sensitive values secure and consider using CI/secrets management for automation contexts.
Common issues:
- If a server fails to start, check that the command and arguments exactly match what your MCP server expects; verify the environment variables are set correctly.
- Ensure network access is permitted for remote endpoints and that firewall settings do not block required ports.
- On macOS, additional permissions may be required for network access or sandboxing depending on how the app is deployed.
Related MCP Servers
gemini-cli
An open-source AI agent that brings the power of Gemini directly into your terminal.
Dive
Dive is an open-source MCP Host Desktop Application that seamlessly integrates with any LLMs supporting function calling capabilities. ✨
lunar
lunar.dev: Agent native MCP Gateway for governance and security
agent
A unified Model Context Protocol server implementation that aggregates multiple MCP servers into one.
mcp-lite-dev
共学《MCP极简开发》项目代码
mcp-client-example
Learn how to implement MCP client with SwiftUI and Anthropic API