Strut

🚀 Deploy Docker Stacks. No Drama.

Strut is an open-source CLI for managing Docker Compose stacks across any VPS that speaks SSH. One command to release, one config tree per project, zero Kubernetes.

What is Strut?

Strut started as a single helper script for managing one client’s VPS. It grew into a full CLI toolkit when I realized the same patterns kept repeating across projects: SSH in, pull the repo, run migrations, restart containers, verify health, backup the database, rotate a key. Same dance, different stack, every time.

It now powers two of my personal stacks and 10+ client stacks across multiple VPSes. Every command is config-driven (no hardcoded service names), every destructive action supports --dry-run, and every release is reproducible. If your infrastructure has SSH and Docker, Strut can ship to it.

✨ Key Capabilities

🧱 Multi-Stack Management

Run as many Docker Compose stacks as you want on a single VPS. Each one gets its own config, secrets, and lifecycle, and they don’t step on each other.

🚀 One-Command Releases

strut my-app release --env prod syncs the repo, runs migrations, deploys containers, and verifies health checks. If anything fails, it tells you exactly where.

🏗️ Two-Tree Architecture

The engine lives at ~/.strut/ and is shared across every project. Your config – strut.conf, stack definitions, env files – lives in the project root. Zero hardcoded values in the engine means you can upgrade across all your machines without touching a single project, and scaffold a new one with strut init in under a minute.

💾 Database Backup & Restore

Native support for Postgres, MySQL, and SQLite. Backup from prod, restore to staging, pull live data down to local for debugging – all from the same CLI.

🔍 Drift Detection

Compare your local config to what’s actually running on the VPS. See exactly what’s different and decide whether to fix it locally or push the change up.

🔐 Secret Rotation

Rotate SSH keys, API tokens, database passwords, and GitHub deploy keys with one command. Strut updates the right .env files and restarts the right services.

🧪 Dry-Run Everywhere

Every destructive command supports --dry-run. See exactly what would happen before it does. Especially useful when you’re rotating credentials at 11pm.

🗺️ VPS Audit & Migration

Already have Docker containers running on a VPS without Strut? strut audit scans the box and helps you convert what’s there into managed stacks, no rip-and-replace required.

🤖 Agent-First by Design

Strut ships with agent steering files and a skills library so AI agents like Claude, Cursor, or any MCP-compatible assistant can run deployments, backups, and rollbacks for you.

I built it this way because that’s how I work now. I describe what I want, the agent reads strut.conf and services.conf to figure out what’s deployed, and runs the right commands. The agent gets context, I get back hours.

• Agent steering: prompt files that guide an agent through Strut commands
• Skills library: pre-built skills for deploy, backup, rollback, scale, logs, and more
• Context-aware: agents read your config and current stack state before acting
• MCP-compatible: works with Claude, Cursor, Windsurf, and any other MCP-compatible agent

🎯 Who It’s For

Solo developers

You have a VPS, a Docker Compose file, and zero interest in learning Kubernetes. Strut gives you a real deployment workflow without the YAML sprawl.

Small teams

Everyone gets the same commands, same configs, same process. New devs can deploy on day one without learning a custom toolchain.

Agencies and consultancies

Same pattern across every client, different configs per project. Stacks stay isolated, the workflow stays consistent, and onboarding a new project takes minutes.

⚡ Quick Start

Install Strut with a single curl:

curl -fsSL https://raw.githubusercontent.com/gfargo/strut/main/install.sh | bash

curl -fsSL https://raw.githubusercontent.com/gfargo/strut/main/install.sh | bash

Then initialize a project, scaffold a stack, and ship it:

strut init --registry ghcr --org my-org       # Initialize project
strut scaffold my-app                          # Create a stack
nano stacks/my-app/.env.template               # Configure (copy to .prod.env)
strut my-app release --env prod                # Deploy to VPS

strut init --registry ghcr --org my-org       # Initialize project
strut scaffold my-app                          # Create a stack
nano stacks/my-app/.env.template               # Configure (copy to .prod.env)
strut my-app release --env prod                # Deploy to VPS

Full walkthrough in the Quick Start guide.

Common day-to-day commands once you’re set up:

strut my-app drift --env prod                    # what's different from local?
strut my-app db:pull --env prod                  # pull prod db down to local
strut my-app keys db:rotate postgres --env prod  # rotate db credentials
strut my-app logs api --follow --env prod        # tail logs from one service
strut my-app health --env prod --json            # health check, machine-readable

strut my-app drift --env prod                    # what's different from local?
strut my-app db:pull --env prod                  # pull prod db down to local
strut my-app keys db:rotate postgres --env prod  # rotate db credentials
strut my-app logs api --follow --env prod        # tail logs from one service
strut my-app health --env prod --json            # health check, machine-readable

📊 Real-World Usage

• 2 personal stacks running production workloads on my own VPSes
• 10+ client stacks deployed across multiple VPS providers
• 21 releases shipped as of v0.20.1, with v1.0 still in active development
• 33 wiki pages covering installation, architecture, deployment workflows, and contribution

🆚 How It Compares

• vs Kubernetes: minutes to first deploy instead of days; no YAML sprawl, no control plane to maintain
• vs Kamal: built-in multi-stack support, drift detection, secret rotation, and database backup workflows
• vs manual SSH: same low learning curve, but with health checks, dry-runs, and a real release pipeline

If your infrastructure runs on a Linux box you can SSH into, Strut probably fits. No proprietary agents, no cloud lock-in, just SSH + Bash + Docker.

🧰 Technology Stack

• Language: Bash (99.7% of the codebase)
• Architecture: Two-tree design – engine at ~/.strut/, your config at project root
• Container runtime: Docker + Docker Compose
• Transport: SSH (no daemons, no agents)
• Registries supported: GitHub Container Registry, Docker Hub, Amazon ECR, or none
• Testing: Bats (Bash Automated Testing System)
• License: MIT

🔗 Links

• Website: strut.griffen.codes
• GitHub: github.com/gfargo/strut
• Wiki / Full Documentation: github.com/gfargo/strut/wiki
• Releases: 21 releases and counting
• Discord: Join the community

Try it on a fresh VPS in five minutes, or point an agent at it and let the agent take the first deploy. Either way, no Kubernetes required.