mcproc MCP server
Commit 5 Okt 2026Memungkinkan AI agent untuk mengelola, memantau, dan melakukan debug pada proses background yang berjalan lama seperti development server melalui CLI dan interface MCP yang terpadu.
Juga dikenal sebagai neptaco/mcproc
- Stars
- 10
- Tools
- 7
- Unduhan, 30 hari
- —
- Versi terbaru
- —
- Commit terakhir
- 5 Okt 2026
- Lisensi
- MIT
Apa itu mcproc MCP?
Memungkinkan AI agent untuk mengelola, memantau, dan melakukan debug pada proses background yang berjalan lama seperti development server melalui CLI dan interface MCP yang terpadu. Server ini menyediakan 7 tool, seperti start_process, stop_process, restart_process.
Ringkasan dibuat dari README proyek
Tentang mcproc MCP server
mcproc menyediakan cara bagi AI agent untuk menangani tugas background yang berjalan lama seperti development server atau build watchers. Tool ini menjembatani celah antara agent otomatis dan workflow terminal manual dengan menawarkan sistem kontrol yang terpadu. Developer dapat menggunakan command line untuk memantau atau mengontrol proses yang sama yang telah dimulai oleh AI agent, memastikan Anda tidak akan terkunci dari environment Anda sendiri.
Sistem ini beroperasi melalui daemon yang ringan, command-line interface, dan MCP server. Untuk menggunakannya, Anda harus mendaftarkan server tersebut ke MCP client seperti Claude Code atau menambahkannya ke file konfigurasi Anda. mcproc mendukung version manager seperti nvm, mise, asdf, dan rbenv untuk mengeksekusi perintah di dalam toolchain tertentu.
SKILL utama yang tersedia untuk agent meliputi start_process, stop_process, dan search_process_logs. Server ini mengelola log menggunakan pola regex dan filtering berbasis waktu, serta mengikuti spesifikasi XDG untuk pengorganisasian file. mcproc juga memungkinkan pengelompokan berbasis proyek dan dapat menunggu pola log tertentu untuk mengonfirmasi bahwa sebuah proses sudah siap.
Apa yang bisa Anda lakukan dengannya
- 1Pemantauan proses secara real-timeDeveloper dapat melihat output langsung dari sebuah proses yang awalnya dimulai oleh AI assistant.
- 2Pembersihan proyek secara massalAnda dapat menghentikan setiap proses yang sedang berjalan yang terkait dengan proyek tertentu hanya dengan satu perintah.
Ditulis berdasarkan README proyek. Setiap use case diperiksa terhadap kutipan dari README tersebut.
Tool mcproc MCP 7
Terdeteksi dari 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 | — |
Konfigurasi
RUST_LOG
| Nama | Deskripsi | Tipe |
|---|---|---|
| RUST_LOG | — |
Pengecekan otomatis
Berjalan 14 Nov 2026Laporan lengkap →Belum diverifikasi. Sebuah server memerlukan skor 80, tidak ada temuan serius, dan data yang cukup untuk kami periksa. Cara kerja checks · Yang perlu diperbaiki untuk menambah poin
- Source repository terhubung dan dapat dijangkauneptaco/mcproc is public and reachablegithub.com/neptaco/mcproc ↗Lolos
- Nama paket mencantumkan repositori ini sebagai sumbernyaTidak diperiksa
- Provenance build (atestasi Sigstore)Tidak diperiksa
- Lisensi open source (OSI)MITspdx.org/licenses/MIT.html ↗Lolos
- Ada commit atau release dalam satu tahun terakhirLast commit or release 2026-10-05github.com/neptaco/mcproc ↗Lolos
- OpenSSF ScorecardTidak diperiksa
- Nama package tidak mirip dengan package populerTidak diperiksa
- Tidak ada advisory kritis pada paket atau dependensinyaTidak diperiksa
- Tidak ada instruksi tersembunyi untuk model7 tool descriptions and the README read; no override phrases, credential paths or hidden charactersLolos
- Tools yang bersifat destruktif telah diberi labelNo tool named like delete, drop or execLolos
- Endpoint remote merespons handshake MCPTidak diperiksa
README
Lihat di 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.
Pertanyaan tentang mcproc MCP
Tool apa yang disediakan mcproc MCP?
Server ini menyediakan 7 tool, termasuk start_process, stop_process, restart_process, list_processes, get_process_logs.
Bagaimana cara menggunakan mcproc dengan Claude Code?
Anda dapat mendaftarkannya sebagai MCP server dengan menjalankan perintah: claude mcp add mcproc mcproc mcp serve.
Bisakah saya memantau proses yang dimulai oleh AI agent?
Ya, proses yang dimulai oleh AI agent dapat diakses sepenuhnya melalui CLI, sehingga Anda bisa memantau log secara real-time.
Apakah mcproc mendukung version manager?
Ya, mcproc mendukung eksekusi perintah melalui berbagai version manager seperti mise, asdf, nvm, dan rbenv.