mcproc

mcproc is an MCP server that bridges the gap between AI agent development and local command-line workflows by managing background processes. It connects AI coding assistants directly to an underlying background daemon and local process environments, allowing developers and autonomous agents to maintain long-running tasks such as local web servers, API daemons, and build watchers. Typically used by software engineers working with AI agents like Claude Code, mcproc enables agents to execute background shell commands, inspect operational state, and parse runtime logs without locking terminal sessions. The server organizes running processes by project context, captures standard output, and supports pattern matching to confirm service readiness. In addition to agent-level controls, developers retain full terminal visibility through an integrated command-line interface to monitor, restart, or terminate tasks. This architecture ensures that AI agents can troubleshoot server crashes, read historical logs, and manage development stacks without losing state between separate prompts.

Category: Cloud & Infrastructure

Tags: automation, background-tasks, process-management, shell

Visit mcproc

How to install and configure mcproc

1. Install mcproc using Homebrew by running brew tap neptaco/tap followed by brew install mcproc, or build from source using cargo build --release after installing protobuf. 2. Start the daemon if not already running with mcproc daemon start. 3. For Claude Code, add the MCP server using the CLI command: claude mcp add mcproc mcproc mcp serve. 4. For other MCP clients like Claude Desktop or Cursor, add the server to your configuration file: json { "mcpServers": { "mcproc": { "command": "mcproc", "args": ["mcp", "serve"] } } } 5. Restart your MCP client to load the process management tools.

What you can do with mcproc

• Launching local web servers or API backends in the background while waiting for specific log readiness patterns. - Searching active or historical process logs with regular expressions to diagnose runtime application errors. - Inspecting the live status and execution metrics of running development tasks across project directories. - Terminating individual stalled background services or performing a clean sweep of all processes across a project. - Running build watchers using local version managers like mise, asdf, nvm, or rbenv without agent terminal lockups.

Key facts

• https://github.com/neptaco/mcproc
• Cloud & Infrastructure, Developer Tools & Code Intelligence
• automation, background-tasks, process-management, shell

Part of MCP Servers

How do I install mcproc?

You can install mcproc on macOS and Linux using Homebrew by tapping neptaco/tap and running brew install mcproc. Alternatively, you can build from source using Rust's cargo build command after installing the protobuf compiler.

What can mcproc do?

mcproc allows AI agents to start, stop, restart, and monitor background processes like local servers or compilers. It captures standard output into searchable logs, supports log pattern readiness checks, and provides a CLI for developers to monitor agent-managed tasks.

Which MCP clients work with mcproc?

mcproc works with Claude Code via its built-in MCP command-line tool, as well as any standard MCP client such as Claude Desktop, Cursor, or custom agents configured via standard input/output execution.

Is mcproc open source?

Yes, mcproc is open-source software released under the MIT License, with its source code publicly available on GitHub.