Deploying personal projects often feels like navigating a minefield. For developers, the allure of free-tier services like Vercel or Netlify for frontends—or Render and Northflank for backends—quickly fades when builds fail, resources throttle, or platforms deprecate features. As one engineer shared in a recent blog post, this frustration led them to design a self-hosted workflow that prioritizes stability and autonomy. Their solution? A minimalist yet robust setup that sidesteps the pitfalls of ephemeral cloud services while remaining accessible for experimental apps.

The Core Workflow: DIY Deployment Simplified

At the heart of this approach is a bare-bones DigitalOcean droplet, the cheapest tier available. While its limited resources rule out on-server Docker builds (which often cause freezes), it excels as a deployment target. The secret sauce? Offloading heavy lifting to GitHub Actions for image builds, pushing artifacts to GitHub Container Registry (GHCR) for efficiency. Deployment is manual but straightforward: images are pulled to the droplet via GitHub workflows, avoiding the complexity of webhooks or tunnels.

The self-hosting platform Dokploy serves as the control center, handling deployments and automating HTTPS certificates through Let's Encrypt for custom domains. As the developer notes:

"Dokploy felt simpler and more consistent than alternatives like Coolify. It abstracts away certificate headaches, letting me focus on the code."

Domain management is streamlined using DigitalOcean’s DNS for subdomains, centralizing what was once scattered across providers like Namecheap. This end-to-end flow—build externally, deploy internally—ensures projects stay online reliably, even for near-zero-traffic demos or proofs-of-concept.

Why This Balance Works for Developers

This setup isn’t about chasing full automation; it’s about predictable control. For personal projects, the overhead of managing a droplet and manual deployments is minimal compared to the anxiety of free-tier uncertainties. The developer emphasizes that knowing every component—from DNS records to build pipelines—reduces debugging nightmares and fosters deeper system understanding. Crucially, it’s cost-effective: a $5/month droplet and open-source tools like Dokploy keep expenses negligible while eliminating vendor lock-in.

In a landscape where platforms pivot or vanish, this self-hosted model offers a sanctuary for experimentation. It’s a reminder that sometimes, the best infrastructure is the one you can troubleshoot at 2 a.m. without a support ticket—proving that for personal work, simplicity isn’t a compromise, but a superpower.

This article is based on a blog post by Bimal on bimals.net. Read the original: Self-Hosting My Projects: A Practical Deployment Setup.