No Unnecessary Deploy Infrastructure with Bare Deploy
Introduction
Software deployment has gotten complicated. What used to be a simple SSH and a few commands has turned into container registries, orchestration layers, image builds, and a growing list of infrastructure to manage. For many of us, that complexity is not solving the problems we actually have.
That is why we are launching Use Bare, an initiative built around a simple idea: software should be free, open-source, self-hosted, and free of unnecessary infrastructure.
Free software. Self-hosted. No unnecessary infrastructure.
It is a philosophy that values clarity, control, and reliability over abstraction for abstraction's sake. And today, we are proud to introduce its first open-source project: Bare Deploy.
The Problem: Deployment Has Become Infrastructure Engineering
Most modern deployment tools assume you need containers, registries, and orchestration. While those tools have their place at massive scale, they come with real trade-offs for smaller teams: more moving parts, more things to debug at 3 AM, and more operational overhead that pulls focus away from building your product.
If you are running a small SaaS, a side project, or a static site on a VPS, do you really need a full container stack? Often, the answer is no. What you need is a simple, reliable way to ship code with zero downtime and the ability to roll back when things go wrong.
Introducing Bare Deploy
Bare Deploy is an atomic, zero-downtime deployment tool for bare VPS servers over SSH. No containers. No agents. No orchestration layers. Just disciplined releases.
Works out-of-the-box. Just disciplined releases. Bare Deploy is designed for developers who want to deploy static websites, Node.js apps, or PHP applications directly to their servers, while keeping full control over the host environment.
How It Works
At its core, Bare Deploy uses a tried-and-true approach: atomic symlink-based releases. Here is how a typical deployment flows:
Bump the version in package.json (optional, but recommended)
Run pre-deploy scripts locally to build or prepare your app
Create a build artifact from the source-code
Acquire a lock on the server to prevent concurrent deploys
Transfer the package to the server via SCP
Extract it into a timestamped release directory
Atomically switch the current symlink to point to the new release
Run post-deploy scripts (like restarting PM2 or clearing caches)
Validate with a health check and auto-rollback if it fails
Release the lock and done
If something goes wrong, manually rolling back is as simple as re-pointing the symlink to a previous release. No rebuilding images. No registry rollbacks. Just instant recovery.
Key Features
Bare Deploy focuses on the essentials that matter for production reliability:
Atomic zero-downtime cutovers with symlink-based releases
Instant rollback to any previous release
Parallel or sequential multi-server deploys (--sequential for sensitive rollouts)
Per-server configuration (different distDir, paths, or settings per host)
SSH key authentication for secure, passwordless deployments
Pre and post-deploy hooks to customize your workflow
Health check validation with automatic rollback on failure
Lock file protection to prevent concurrent deployments
Dry-run mode to preview changes before going live
Release pruning (bare cleanup) to keep only the last N releases
PHP support with OPcache fixes (opcache.revalidate_path=1 via .user.ini)
Static site handling that preserves .well-known/ directories (perfect for Let's Encrypt)
Git-aware packaging that respects .gitignore while including build outputs
Why No Containers?
Containers solve real problems at scale. But they also introduce registry management, image lifecycles, and orchestration complexity that many teams simply do not need.
Bare Deploy keeps things aligned with the host filesystem and your process manager (like PM2, systemd, or Supervisor). That means less abstraction, faster debugging, and operational clarity. It is a great fit for early-stage SaaS, indie projects, and small teams that value control and reliability over scale for scale's sake.
Quick Start
Getting started takes just a few commands:
# Install globally
npm install -g bare-deploy
# Initialize configuration in your project
bare init
# Deploy to your configured servers
bare deploy
From there, you can use bare list to see releases, bare rollback [id] to roll back, and bare cleanup to prune old releases. Check out the full documentation for all available flags and options.
What Comes Next
The "Use Bare" initiative is just getting started. Bare Deploy is our first step toward a catalog of tools that embrace simplicity, self-hosting, and freedom from unnecessary infrastructure.
We built this because we believe you should own your stack. No lock-in. No forced subscriptions. No layers of complexity you do not need.
If this resonates with you, we would love your feedback, ideas, or contributions. Head over to the GitHub repository to star it, open an issue, or submit a pull request.
Free software. Self-hosted. No unnecessary infrastructure.
Bare Deploy is built by Abtz Labs, a software studio based in Tokyo, Japan, creating practical tools for developers since 2018.
About the Author
Rogerio is a seasoned Full-Stack Engineer, Product Owner, and Entrepreneur with a deep background in Information Systems. He bridges the gap between complex technical stacks and a customer-centric, product-first mindset.