mcproc MCP server
Commit Oct 5, 2026Enables AI agents to manage, monitor, and debug long-running background processes like development servers through a unified CLI and MCP interface.
Also known as neptaco/mcproc
- Stars
- 10
- Tools
- 7
- Downloads, 30 days
- —
- Latest version
- —
- Last commit
- Oct 5, 2026
- License
- MIT
What is mcproc MCP?
mcproc enables AI agents to manage, monitor, and debug long-running background processes like development servers through a unified CLI and MCP interface. It exposes 7 tools, such as start_process, stop_process, restart_process.
Summary generated from the project's README
About mcproc MCP server
mcproc provides a way for AI agents to handle long-running background tasks like development servers or build watchers. It bridges the gap between automated agents and manual terminal workflows by offering a unified control system. Developers can use the command line to monitor or control the same processes that an AI agent has started, ensuring you are never locked out of your own environment.
The system operates through a lightweight daemon, a command-line interface, and an MCP server. To use it, you must register the server with an MCP client like Claude Code or add it to your configuration file. It supports version managers such as nvm, mise, asdf, and rbenv to execute commands within specific toolchains.
Key tools available to agents include start_process, stop_process, and search_process_logs. The server manages logs using regex patterns and time-based filtering, and it follows XDG specifications for file organization. It also allows for project-aware grouping and can wait for specific log patterns to confirm a process is ready.
What you can do with it
- 1Real-time process monitoringA developer can watch the live output of a process that was originally started by an AI assistant.
- 2Bulk project cleanupYou can stop every running process associated with a specific project using a single command.
Written from the project's README. Every use case is checked against a quote from it.
mcproc MCP tools 7
Detected from the READMEstart_processstop_processrestart_processlist_processesget_process_logssearch_process_logsget_process_status
| start_process | — | |
| stop_process | — | |
| restart_process | — | |
| list_processes | — | |
| get_process_logs | — | |
| search_process_logs | — | |
| get_process_status | — |
Configuration
RUST_LOG
| Name | Description | |
|---|---|---|
| RUST_LOG | — |
Automated checks
Ran Nov 14, 2026Full report →Not verified yet. A server needs a score of 80, no serious finding, and enough data for us to check. How checks work · What to fix for more points
- Source repository linked and reachableneptaco/mcproc is public and reachablegithub.com/neptaco/mcproc ↗Pass
- Package names this repository as its sourceNot checked
- Build provenance (Sigstore attestation)Not checked
- Open source license (OSI)MITspdx.org/licenses/MIT.html ↗Pass
- Commit or release in the last yearLast commit or release 2026-10-05github.com/neptaco/mcproc ↗Pass
- OpenSSF ScorecardNot checked
- Package name not close to a popular packageNot checked
- No critical advisory in the package or its dependenciesNot checked
- No hidden instructions to the model7 tool descriptions and the README read; no override phrases, credential paths or hidden charactersPass
- Destructive tools are labelledNo tool named like delete, drop or execPass
- Remote endpoint answers the MCP handshakeNot checked
README
View on GitHubmcproc
A Model Context Protocol (MCP) server for comfortable background process management on AI agents.
Overview
mcproc bridges the gap between AI agent development and traditional command-line workflows. It enables AI agents to manage long-running development processes (like dev servers, build watchers, etc.) while providing developers with full CLI access to monitor and control these same processes.
Why mcproc?
Simple AI agent-launched processes are stateless and can't manage long-running processes effectively. mcproc solves this by:
- Unified Control: No more confusion about which agent or terminal is running what - all processes are centrally managed
- Context Preservation: Logs are captured and stored, allowing AI agents to debug issues while reviewing logs from earlier
- Developer-Friendly: Full CLI access means you're never locked out of your own development environment
Key Features
- 🔄 Unified Process Management: Start and manage background processes from AI agents via MCP, then monitor them from your terminal
- 👁️ Cross-Environment Visibility: Processes started by AI agents are fully accessible via CLI and other agents, and vice versa
- 📝 Intelligent Log Management: Capture, persist, and search process logs with powerful regex patterns
- 📁 Project-Aware: Automatically groups processes by project context
- 📊 Real-time Monitoring: Follow logs in real-time from CLI while AI agents manage the processes
- 🛡️ XDG Compliant: Follows XDG Base Directory specification for proper file organization
- ⚡ Wait-for-Log: Start processes and wait for specific log patterns to ensure readiness
- 🔍 Advanced Search: Time-based filtering, context lines, and regex support for log analysis
- 🧰 Toolchain Support: Execute commands through version managers (mise, asdf, nvm, rbenv, etc.)
- 🧹 Clean Command: Stop all processes in a project with a single command
- 🌲 Process Groups: Automatic cleanup of child processes when stopping parent processes
Installation
Using Homebrew (macOS and Linux)
# Add the tap
brew tap neptaco/tap
# Install mcproc
brew install mcproc
Build from source
Prerequisites
- Rust toolchain (rustc, cargo)
- protobuf compiler:
- macOS:
brew install protobuf - Linux:
apt-get install protobuf-compiler
- macOS:
git clone https://github.com/neptaco/mcproc.git
cd mcproc
cargo build --release
# Install to PATH (optional)
cargo install --path mcproc
Usage
Setup as MCP Server
After installing mcproc, you need to register it as an MCP server with your AI assistant.
For Claude Code
# Register mcproc as an MCP server
claude mcp add mcproc mcproc mcp serve
For Other MCP Clients
Configure your MCP client by adding mcproc to your configuration:
{
"mcpServers": {
"mcproc": {
"command": "mcproc",
"args": ["mcp", "serve"]
}
}
}
Available MCP Tools
Once registered, AI agents can use these tools:
start_process: Start a development server or background processstop_process: Stop a running processrestart_process: Restart a processlist_processes: List all running processesget_process_logs: Retrieve process logssearch_process_logs: Search through process logs with pattern matchingget_process_status: Get detailed process information
For Developers (CLI)
While AI agents manage processes in the background, you can monitor and control them:
Recommended command: mcproc logs -f
CLI Commands
| Command | Description | Flags | Example |
|---|---|---|---|
🗒️ ps |
List all running processes | -s, --status <STATUS> Filter by status |
mcproc ps --status running |
🚀 start **<NAME>** |
Start a new process | -c, --cmd <CMD> Command to run-d, --cwd <DIR> Working directory-e, --env <KEY=VAL> Environment variables-p, --project <NAME> Project name--wait-for-log <PATTERN> Wait for log pattern--wait-timeout <SECS> Wait timeout--toolchain <TOOL> Version manager to use |
mcproc start web -c "npm run dev" -d ./app |
🛑 stop **<NAME>** |
Stop a running process | -p, --project <NAME> Project name-f, --force Force kill (SIGKILL) |
mcproc stop web -p myapp |
🔄 restart **<NAME>** |
Restart a process | -p, --project <NAME> Project name |
mcproc restart web |
📜 logs **<NAME>** |
View process logs | -p, --project <NAME> Project name-f, --follow Follow log output-t, --tail <NUM> Number of lines to show |
mcproc logs web -f -t 100 |
🔍 grep **<NAME>** **<PATTERN>** |
Search logs with regex | -p, --project <NAME> Project name-C, --context <NUM> Context lines-B, --before <NUM> Lines before match-A, --after <NUM> Lines after match--since <TIME> Search since time--until <TIME> Search until time--last <DURATION> Search last duration |
mcproc grep web "error" -C 3 |
🧹 clean |
Stop all processes in project | -p, --project <NAME> Project name-f, --force Force kill |
mcproc clean -p myapp |
🎛️ daemon start |
Start mcproc daemon | None | mcproc daemon start |
🎛️ daemon stop |
Stop mcproc daemon | None | mcproc daemon stop |
🎛️ daemon status |
Check daemon status | None | mcproc daemon status |
🔌 mcp serve |
Run as MCP server | None | mcproc mcp serve |
ℹ️ --version |
Show version info | None | mcproc --version |
❓ --help |
Show help message | None | mcproc --help |
Examples
# Start the daemon (if not already running)
mcproc daemon start
# View all processes (including those started by AI agents)
mcproc ps
# Follow logs in real-time
mcproc logs frontend -f
# Multi-process log streaming per project
mcproc logs -f
# Search through logs
mcproc grep backend "error" -C 5
# Stop a process
mcproc stop frontend
Example Workflow
-
AI agent starts your development server:
Agent: "I'll start the frontend dev server for you" → Uses MCP tool: start_process(name: "frontend", cmd: "npm run dev", wait_for_log: "Server running") -
You monitor it from terminal:
mcproc logs -f # See real-time logs as the server runs -
AI agent detects an error and searches logs:
Agent: "Let me check what's causing that error" → Uses MCP tool: search_process_logs(name: "frontend", pattern: "ERROR|WARN", last: "5m") -
You can see the same information:
mcproc grep frontend "ERROR|WARN" -C 3 --last 5m
Advanced Examples
# Start a process with environment variables
mcproc start api --cmd "python app.py" --env PORT=8000 --env DEBUG=true
# Wait for a specific log pattern before considering the process ready
mcproc start web --cmd "npm run dev" --wait-for-log "Server running on" --wait-timeout 60
# Search logs with time filters
mcproc grep api "database.*connection" --since "14:30" --until "15:00"
# View logs from multiple processes in the same project
mcproc ps
mcproc logs web --project myapp -t 100
# Use version managers for Node.js projects
mcproc start web --cmd "npm run dev" --toolchain nvm
mcproc start api --cmd "yarn start" --toolchain mise
# Clean up all processes in a project
mcproc clean --project myapp
# Force stop all processes in current project
mcproc clean --force
Architecture
mcproc consists of three main components:
- mcproc daemon: A lightweight daemon that manages processes and handles log persistence
- mcproc CLI: Command-line interface for developers to interact with the daemon
- MCP Server: Exposes process management capabilities to AI agents via the Model Context Protocol
File Locations (XDG Compliant)
- Config:
$XDG_CONFIG_HOME/mcproc/config.toml(defaults to~/.config/mcproc/) - Logs:
$XDG_STATE_HOME/mcproc/log/(defaults to~/.local/state/mcproc/log/) - Runtime:
$XDG_RUNTIME_DIR/mcproc/(defaults to/tmp/mcproc-$UID/)
Development
Building from Source
# Clone the repository
git clone https://github.com/neptaco/mcproc.git
cd mcproc
# Build all components
cargo build --release
# Run tests
cargo test
# Run with verbose logging
RUST_LOG=mcproc=debug cargo run -- daemon start
Project Structure
mcproc/
├── mcproc/ # CLI and daemon implementation
├── mcp-rs/ # Reusable MCP server library
├── proto/ # Protocol buffer definitions
└── docs/ # Architecture and design documentation
Contributing
Contributions are welcome! Please feel free to submit a Pull Request.
License
MIT License
Copyright (c) 2025 Atsuhito Machida (neptaco)
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
mcproc MCP questions
What tools does mcproc MCP provide?
It exposes 7 tools, including start_process, stop_process, restart_process, list_processes, get_process_logs.
How can I use mcproc with Claude Code?
You can register it as an MCP server by running the command: claude mcp add mcproc mcproc mcp serve.
Can I monitor processes started by an AI agent?
Yes, processes started by AI agents are fully accessible via the CLI, allowing you to follow logs in real-time.
Does mcproc support version managers?
Yes, it supports executing commands through various version managers such as mise, asdf, nvm, and rbenv.