Self-Hosted Cloud IDE: code-server, Coder & DevPod on Your Own VPS (2026)

GitHub Codespaces charges $0.18 per core-hour. Sounds harmless — until the bill arrives: At 8 hours per day and 20 work days, that's $28.80 per month per developer. For a 2-core setup. Need 4 cores for TypeScript projects? Double the price to $57.60.

A VPS with code-server? From €4 per month. Unlimited hours, full VS Code extensions, root access. And your code stays on your server instead of on Microsoft Azure.

This guide covers the four best self-hosted cloud IDEs, how to install them in under 10 minutes, and when self-hosting makes financial sense.

Why Self-Host a Cloud IDE?

The Cost Problem

Cloud IDEs are convenient — but expensive. GitHub Codespaces starts with 120 free core-hours per month, but those are used up within days of full-time usage. After that, you pay per hour. Gitpod Classic was discontinued in October 2024; its successor Gitpod Flex is only available as a self-hosted solution, requiring your own infrastructure anyway. Replit costs $25/month for Pro features.

For a 5-person team, Codespaces quickly adds up to $150–300 per month — for something you can cover on a single VPS for €10–15.

Data Sovereignty and GDPR Compliance

For European developers and companies, there's another argument: With GitHub Codespaces, your code runs on Azure servers in the US. If you're working on client projects with personal data, this can be a GDPR problem.

With a self-hosted cloud IDE on a European VPS (Hetzner, IONOS, Netcup), your code stays in Germany. No data processing agreements with US clouds needed, no discussions with the data protection officer.

No Vendor Lock-in

You're not tied to any provider. Switch VPS providers? Done in an hour — copy Docker Compose, migrate data, update DNS. Migrate away from Codespaces? That takes days: rewriting dev container configurations, migrating secrets, adjusting team workflows.

The 4 Best Self-Hosted Cloud IDEs Compared

All four options are based on VS Code — the difference lies in multi-user support, management, and infrastructure complexity.

code-server

code-server by Coder Inc. is the simplest option: One Docker container, one password, VS Code in the browser. Installation takes 5 minutes. Perfect for solo developers who want to access their projects from anywhere — laptop, tablet, or someone else's computer.

Limitation: Only one user at a time. No workspace isolation for teams and no centralized management.

Coder

Coder is the enterprise solution. Each developer gets their own workspace from a Terraform template — reproducible, isolated, centrally managed. Coder supports OAuth/OIDC (GitHub, Google, Okta), has an admin dashboard, and a REST API for automation.

Coder requires PostgreSQL 13+ as backend database. Resource requirements are higher, but for teams of 3+ developers, centralized management justifies the overhead.

OpenVSCode Server

OpenVSCode Server by Gitpod is the leanest candidate: A single binary, ~1 GB RAM, no Docker needed. However, no multi-user support and no built-in authentication. You need a reverse proxy with basic auth or a VPN like Tailscale in front of it.

Ideal for developers who want maximum simplicity and handle security at the network layer.

DevPod

DevPod by Loft Labs takes a different approach: Instead of a central server, DevPod uses devcontainer.json files and launches workspaces via SSH on a VPS, in Docker, or on Kubernetes. The client runs locally, the workspace runs remotely.

Advantage: No running server daemon, workspaces start on-demand. Disadvantage: Higher resource usage per workspace (~2 GB RAM) and a slightly more complex setup.

Which VPS Is Enough? Hardware Requirements

The IDE itself is lightweight — the load comes from language servers (TypeScript, Python, Go), linters, compilers, and Docker containers. Here's a reference:

Important: NVMe SSD makes a noticeable difference in IDE responsiveness. Language servers constantly read from disk — on HDD or slow SSD, you feel it on every auto-complete.

Installation: code-server in 5 Minutes (Docker Compose)

The simplest setup: code-server behind Caddy as reverse proxy with automatic HTTPS.

Prerequisites

• VPS with Ubuntu 22.04/24.04 and Docker installed
• Domain or subdomain (e.g., code.example.com) with DNS pointing to the VPS IP

Docker Compose

Create the file /opt/code-server/docker-compose.yml on the VPS:

services:
  code-server:
    image: codercom/code-server:latest
    container_name: code-server
    restart: unless-stopped
    environment:
      - PASSWORD=your-secure-password
    volumes:
      - code-data:/home/coder
    networks:
      - web

  caddy:
    image: caddy:2-alpine
    container_name: caddy
    restart: unless-stopped
    ports:
      - "80:80"
      - "443:443"
    volumes:
      - ./Caddyfile:/etc/caddy/Caddyfile
      - caddy-data:/data
    networks:
      - web

volumes:
  code-data:
  caddy-data:

networks:
  web:

Create the Caddyfile in the same directory:

code.example.com {
    reverse_proxy code-server:8080
}

Start it:

cd /opt/code-server
docker compose up -d

After 30 seconds, VS Code is available at https://code.example.com. Caddy automatically provisions a Let's Encrypt certificate.

Tip: For Docker workspaces inside code-server, mount the Docker socket: add -v /var/run/docker.sock:/var/run/docker.sock to the code-server service.

Installation: Coder for Teams

Coder needs PostgreSQL as its database. This setup runs entirely in Docker Compose:

services:
  coder:
    image: ghcr.io/coder/coder:latest
    container_name: coder
    restart: unless-stopped
    environment:
      - CODER_ACCESS_URL=https://coder.example.com
      - CODER_PG_CONNECTION_URL=postgresql://coder:coder-password@db:5432/coder?sslmode=disable
    ports:
      - "3000:3000"
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock
    depends_on:
      db:
        condition: service_healthy
    networks:
      - coder

  db:
    image: postgres:16-alpine
    container_name: coder-db
    restart: unless-stopped
    environment:
      - POSTGRES_USER=coder
      - POSTGRES_PASSWORD=coder-password
      - POSTGRES_DB=coder
    volumes:
      - coder-db:/var/lib/postgresql/data
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U coder"]
      interval: 5s
      timeout: 5s
      retries: 5
    networks:
      - coder

volumes:
  coder-db:

networks:
  coder:

After docker compose up -d, the Coder dashboard is available on port 3000. The first user automatically becomes admin.

Workspace Templates

Coder uses Terraform templates to define workspaces. Example for a Docker-based workspace:

coder templates create docker \
  --directory ./templates/docker

Each developer then gets an isolated workspace with preconfigured tools, extensions, and git credentials — with a single click from the dashboard.

Note: For HTTPS in front of Coder, set up a reverse proxy (Caddy or Traefik). The CODER_ACCESS_URL must point to the public URL.

Cost Comparison: Self-Hosted vs. Codespaces vs. Replit

Monthly costs at typical full-time usage (160h/month):

Bottom line: Self-hosting pays off from the first developer. From 3 developers, the difference is dramatic: ~€10 vs. ~$87. The only downside is the one-time setup effort — which you can complete in under 10 minutes with the Docker Compose files above.

Security: HTTPS, Auth, and Network

A cloud IDE on the internet without protection is an open invitation. Minimum requirements:

• HTTPS: Mandatory. Caddy or Traefik with Let's Encrypt solve this automatically.
• Authentication: code-server has built-in password protection. For Coder: set up OAuth/OIDC (GitHub login is configured in 5 minutes).
• Firewall: Only open port 443 (HTTPS) and 22 (SSH). Never expose code-server or Coder directly on port 8080/3000.
• SSH alternative: Instead of browser access, you can use VS Code Remote-SSH and use the VPS as a pure backend server. No reverse proxy needed.
• Zero trust: Tailscale or WireGuard as a VPN layer in front of the IDE — then no port is publicly reachable.

Practical Tips

Language servers need RAM: TypeScript with large monorepos can consume 1–2 GB RAM on its own. Python with Pylance is similar. Plan at least 2 GB per active workspace.

Cache git credentials: Set up git credential-store or an SSH agent so you don't have to enter passwords on every push. With Coder, git credentials can be stored as workspace parameters.

Automate dotfiles: code-server supports a dotfiles repository. On start, .bashrc, .gitconfig, and VS Code settings are automatically applied. Coder has a dedicated dotfiles feature in its dashboard.

Monitoring: Install htop or btop on the VPS to monitor RAM and CPU usage. With Coder, the dashboard shows resource consumption per workspace.

Conclusion

Self-hosted cloud IDEs are no longer a niche tool in 2026. code-server for solo developers, Coder for teams — both are installed in under 10 minutes and cost a fraction of Codespaces or Gitpod.

The only question is: Is the one-time setup effort worth it? For a team of 2+, the answer is clearly yes — you save from the first month. For solo developers, the financial advantage is smaller, but the benefits of data sovereignty and control remain.