March 20, 2026 (3mo ago)
Written by David Viejo
Last updated March 20, 2026 (3mo ago)
You can deploy a Next.js app to a VPS without Vercel by running Node.js + PM2 + Nginx on any $4–6/mo server — or by using a platform like Temps that handles all of that automatically from a single git push. This guide covers both paths so you can choose what fits your situation.
TL;DR: Manual deployment means managing Node.js, PM2, Nginx, Certbot, deploy scripts, and GitHub Actions — plus separate services for monitoring, error tracking, and analytics. A self-hosted deployment platform collapses all of that into one tool. Either way, this guide walks every step.
The fastest answer: SSH into a fresh Ubuntu server, install Node 20, clone your repo, run npm ci && npm run build, start the app with PM2, put Nginx in front for SSL, and wire up GitHub Actions to re-run those steps on every push to main.
The rest of this guide explains exactly how to do that, why each piece exists, and what it still leaves unsolved.
Before diving into the steps, it helps to know what you're signing up for:
| Manual (PM2 + Nginx) | Temps (self-hosted) | Vercel | |
|---|---|---|---|
| Setup time | 30–60 min per app | ~15 min total | Minutes |
| Cost | VPS + your time | VPS + Temps free | Free tier → see pricing page |
| Deployments | Custom deploy.sh + GitHub Actions | git push auto-deploy | git push |
| Zero-downtime | PM2 cluster mode (manual) | Built-in health checks + rollback | Yes |
| SSL | Certbot (manual renewal) | Automatic | Automatic |
| Analytics | Separate tool required | Built-in (privacy-first) | Separate tool |
| Error tracking | Sentry/separate | Built-in | Sentry/separate |
| Rollback | Manual git revert + redeploy | One-click | One-click |
| Self-hosted | Yes | Yes (Apache 2.0) | No |
| Vendor lock-in | None | None | Yes |
Temps is a single Rust binary that provides the git-push deployment workflow plus built-in analytics, error tracking, session replay, uptime monitoring, and managed databases — free to self-host, or ~$6/mo on Temps Cloud (Hetzner cost + 30%, no per-seat fees, no bandwidth bills). If you want the manual approach for full control or learning, keep reading. If you want to skip straight to a platform, see how Temps handles Next.js deployments.
npm run build locally.This guide assumes Ubuntu 22.04 or 24.04. Debian works too with minor differences.
SSH into your new server:
ssh root@your-server-ip
Update packages and install the basics:
apt update && apt upgrade -y
apt install -y curl git ufw
Set up the firewall. Open SSH, HTTP, and HTTPS. Close everything else:
ufw allow OpenSSH
ufw allow 80
ufw allow 443
ufw enable
Create a non-root user (running everything as root is asking for trouble):
adduser deploy
usermod -aG sudo deploy
Copy your SSH key to the new user:
rsync --archive --chown=deploy:deploy ~/.ssh /home/deploy
Log out and log back in as the deploy user from now on.
Don't install Node from apt. The version in Ubuntu's default repos is usually ancient. Use the NodeSource repository:
curl -fsSL https://deb.nodesource.com/setup_20.x | sudo -E bash -
sudo apt install -y nodejs
Verify:
node --version # Should be 20.x
npm --version
If your project uses a specific Node version (check .nvmrc or engines in package.json), match it here. Mismatched Node versions between local and server are one of the most common causes of "works on my machine" deployment failures.
cd /home/deploy
git clone https://github.com/your-username/your-nextjs-app.git
cd your-nextjs-app
npm ci
npm ci instead of npm install. It installs from the lockfile exactly, which is what you want in production. npm install can resolve to different versions.
Set your environment variables:
cp .env.example .env.production
nano .env.production
# Fill in your production values: DATABASE_URL, API keys, etc.
Build:
npm run build
If this fails, fix it before continuing. Common issues: missing env vars that the build needs at compile time (Next.js bakes NEXT_PUBLIC_* variables into the client bundle during build), or native dependencies that need build tools (sudo apt install -y build-essential).
You need a process manager. If you just run npm start in your terminal and disconnect, the process dies. PM2 keeps it running, restarts it if it crashes, and manages logs.
sudo npm install -g pm2
Start your app:
pm2 start npm --name "nextjs-app" -- start
Check it's running:
pm2 status
pm2 logs nextjs-app
By default, Next.js starts on port 3000. Verify it's listening:
curl http://localhost:3000
Tell PM2 to start your app on server boot:
pm2 startup
pm2 save
The pm2 startup command prints a line you need to copy and run with sudo. Don't skip it, or your app won't survive a server reboot.
Nginx sits in front of your Next.js app. It handles SSL termination, serves static files, and proxies dynamic requests to your Node process.
sudo apt install -y nginx
Create a site config:
sudo nano /etc/nginx/sites-available/your-app
server {
listen 80;
server_name yourdomain.com www.yourdomain.com;
location / {
proxy_pass http://127.0.0.1:3000;
proxy_http_version 1.1;
proxy_set_header Upgrade $http_upgrade;
proxy_set_header Connection 'upgrade';
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header X-Forwarded-Proto $scheme;
proxy_cache_bypass $http_upgrade;
}
}
Enable the site:
sudo ln -s /etc/nginx/sites-available/your-app /etc/nginx/sites-enabled/
sudo nginx -t
sudo systemctl reload nginx
The proxy_set_header Upgrade and Connection 'upgrade' lines are for WebSocket support. If your app uses real-time features, these headers are required.
At this point, your app should be accessible at http://yourdomain.com. No SSL yet.
sudo apt install -y certbot python3-certbot-nginx
sudo certbot --nginx -d yourdomain.com -d www.yourdomain.com
Certbot modifies your Nginx config to add SSL, sets up automatic certificate renewal, and redirects HTTP to HTTPS.
Verify the renewal timer is active:
sudo systemctl status certbot.timer
Certificates expire every 90 days. Certbot's timer renews them automatically. If you skip this check and the timer isn't running, your site goes down in 3 months with an expired cert. It happens more often than you'd think.
Your app is live. Now you need a way to update it when you push new code.
Create /home/deploy/deploy.sh:
#!/bin/bash
set -e
cd /home/deploy/your-nextjs-app
echo "Pulling latest code..."
git pull origin main
echo "Installing dependencies..."
npm ci
echo "Building..."
npm run build
echo "Restarting..."
pm2 restart nextjs-app
echo "Done."
chmod +x /home/deploy/deploy.sh
# .github/workflows/deploy.yml
name: Deploy
on:
push:
branches: [main]
jobs:
deploy:
runs-on: ubuntu-latest
steps:
- name: Deploy to VPS
uses: appleboy/ssh-action@v1
with:
host: ${{ secrets.SERVER_IP }}
username: deploy
key: ${{ secrets.SSH_PRIVATE_KEY }}
script: /home/deploy/deploy.sh
Add SERVER_IP and SSH_PRIVATE_KEY to your repo's GitHub Secrets.
The deploy script above has real gaps:
main goes straight to production.You can solve each of these individually (blue-green deploys with Nginx upstream toggling, a webhook server, PM2's cluster mode). But each solution adds complexity, and by the time you've built all of them, you've built a deployment platform. For a deeper look at how these platforms automate this pipeline, see how git-push deployments work under the hood.
Your app is deployed. How do you know it's still running tomorrow?
Logs: PM2 handles application logs. Rotate them or they'll fill your disk:
pm2 install pm2-logrotate
Uptime: You need something that pings your site and alerts you when it's down. Free options: UptimeRobot, Betterstack (free tier), or a cron job that curls your health endpoint. See our guide on building an uptime monitoring system for more depth.
Error tracking: When your app throws an unhandled exception in production, how do you find out? Sentry (free tier), LogRocket, or parsing PM2 logs manually. We covered the options in error tracking without Sentry.
Analytics: Google Analytics, Plausible, Umami, or similar. If you want to avoid third-party scripts entirely, see how to add web analytics without third-party scripts.
Each of these is a separate tool, a separate account, a separate dashboard.
Let's count:
Running on your server:
Still need but didn't set up: 7. Uptime monitoring (external service) 8. Error tracking (external service) 9. Analytics (external service) 10. Log management (PM2 + logrotate, or external service)
That's 6 things on your server and 4 external services for a single Next.js app.
If you'd rather not manage Nginx, PM2, Certbot, and four separate observability tools, Temps is a self-hosted deployment platform that replaces all of it.
Three things Temps does that the manual approach doesn't:
Nixpacks auto-detection. Point Temps at your GitHub repo. It detects next in package.json and builds your app automatically — no Dockerfile, no build scripts to maintain. It detects standalone output mode too, using node server.js instead of npx next start when appropriate.
Health checks with automatic rollback. Temps checks your app every 5 seconds after deploy. It requires 2 consecutive successful responses before marking the deployment live. If the new version fails health checks within the 60-second error window, Temps rolls back to the previous deployment automatically. The manual PM2 approach has no equivalent.
Built-in observability stack. Analytics, error tracking, session replay, and uptime monitoring are included — not separate accounts. Temps replaces PostHog/Plausible, Sentry, FullStory, and Pingdom alongside its deployment engine.
Self-host Temps on the same VPS (Apache 2.0, free):
curl -fsSL https://temps.sh/install.sh | sh
temps setup
Or use Temps Cloud — managed Hetzner servers at ~$6/mo (Hetzner cost + 30%, no per-seat fees, no bandwidth bills). Then connect your Next.js repo and push.
Yes. The manual path in this guide (Node.js + PM2 + Nginx) runs the app directly without Docker. Temps also supports Dockerfile-based builds if you prefer containerization.
No. Temps uses Nixpacks to auto-detect Next.js projects from package.json. If your repo has a next dependency, Temps detects it as a Next.js app and configures the build automatically. A Dockerfile is supported but not required.
2 GB RAM is the practical minimum for a Next.js app that builds on the server. 1 GB servers frequently run out of memory during next build. If you're also running a database on the same box, use 4 GB.
Yes. Temps runs Next.js as a standard Node.js application. App Router, Server Components, API routes, and Server Actions all work as expected — there are no serverless function limitations or Edge runtime constraints.
Temps runs health checks every 5 seconds after a new deployment starts, requiring 2 consecutive successful responses before routing traffic to the new version. If health checks fail within the 60-second error window, it rolls back to the previous version automatically.