Tow

v0.3.1 · MIT

A lightweight, agentless deployment orchestrator. Deploy to bare-metal servers and cloud VMs without Kubernetes.

Deployment DevOps CLI SSH Go Agentless MCP

Quick Install

Homebrew
brew install neurosamAI/tap/tow
npm
npm install -g @neurosamai/tow
Go
go install github.com/neurosamAI/tow-cli/cmd/tow@latest
Binary
curl -fsSL https://tow-cli.neurosam.ai/install.sh | sh

Key Features

Agentless SSH Deployment

No agent needs to be installed on target servers. All you need is SSH to deploy instantly.

Symlink-Based Atomic Deployment

A symlink switch enables instant rollback. Minimizes downtime during deployment.

Project Auto-Detection

A single `tow init` command auto-detects your project type, framework, and build tool, then generates the configuration.

12 Built-in Module Handlers

Supports 12 languages/frameworks out of the box, including Spring Boot, Node.js, Python, Go, and Rust.

35+ Infrastructure Plugins

Built-in plugins for Kafka, Redis, MySQL, PostgreSQL, and more, plus a community plugin ecosystem (tow plugin add).

4 Health Check Types

HTTP, TCP, log pattern, and custom command — four built-in health check methods.

Parallel Execution

Deploy to multiple servers simultaneously. Also supports rolling updates and automatic rollback.

Lifecycle Hooks

Run custom scripts before and after build, deploy, start, and stop.

AI Agent Integration

A built-in MCP server enables native integration with AI agents like Claude, Cursor, and Windsurf.

Deployment Metrics

Use tow metrics to view deployment frequency, action breakdowns, and per-module statistics as bar charts.

Interactive Selection

Omit the server/module flags for an interactive prompt. View logs from multiple modules at once using comma-separated names.

Overview

Quick Start

Tow is a deployment tool that fills the gap between shell scripts and Kubernetes. It provides a simple, reliable deployment pipeline for VM-based infrastructure.

Basic Workflow

# Detect project & generate config file
tow init

# Initialize remote server
tow setup -e prod -m api-server

# One-click deploy (build → package → upload → deploy)
tow auto -e prod -m api-server

# Check status
tow status -e prod -m api-server

# Instant rollback
tow rollback -e prod -m api-server

Main Commands

CommandDescription
tow initAuto-detect project & generate config
tow autoRun the full pipeline
tow deployPackage → upload → install → restart
tow statusCheck module status (PID, uptime, memory)
tow rollbackInstantly restore the previous deployment
tow logsStream remote logs (--all, -F, presets)
tow sshRun ad-hoc commands on remote servers
tow diffCompare local vs remote code before deploying
tow configManage server/module configuration via CLI
tow metricsView deployment frequency and per-action/module statistics
tow pluginInstall/remove community plugins
tow provisionProvision a server (timezone, JRE, tools)
tow mcp-serverStart the MCP server for AI agent integration

Comparison

TowAnsibleCapistranoKamal
InstallSingle binaryRequires PythonRequires RubyRequires Docker
Agentless
No Docker needed
Project auto-detection
Built-in health checks4 typesPlugin
Instant rollback✓ (symlink)Re-run needed✓ (symlink)✓ (container)
Multi-language support12 typesPlaybooksRuby-centricDocker images
Multi-server log streaming✓ (color-coded)
Pre-deploy diff comparison
AI agent integration (MCP)
Deployment metrics
Interactive selection
Community pluginsGalaxy

Try it out

Feedback and contributions are welcome — open an issue or send a PR on GitHub.