The platform workflow without platform pricing
Usage-based platforms are pleasant to deploy to and hard to budget for. Every vCPU, every gigabyte of RAM and every gigabyte of traffic is metered, and a JVM that keeps its heap warm all day is exactly the kind of workload that makes the bill grow. Here you get the same push-to-deploy workflow on a VPS that only runs your apps, for a fixed monthly price.
- Git push deploys. Connect GitHub, GitLab or any Git repository. Every push builds and deploys, with preview deployments for pull requests.
- Builds Java the way you already do. Use your own Dockerfile, Docker Compose, or let Nixpacks detect Maven and Gradle projects.
- Databases in one click. PostgreSQL, MySQL, MariaDB, MongoDB and Redis, with scheduled backups to S3-compatible storage.
- Automatic HTTPS. Point a domain at the server and Coolify issues and renews Let's Encrypt certificates for every app.
- An API for your coding agent. Create a token and Claude Code, OpenCode or your CI pipeline can create and deploy apps for you.
- No lock-in. Coolify is open source and your apps are plain containers.
How a deploy works
Add a Dockerfile to your project, or let Coolify detect the build. For a Spring Boot JAR, this is enough:
FROM eclipse-temurin:21-jre
COPY target/app.jar /app.jar
ENV JAVA_TOOL_OPTIONS="-XX:MaxRAMPercentage=60"
EXPOSE 8080
ENTRYPOINT ["java", "-jar", "/app.jar"]
In Coolify, create a new resource from your repository, set the port to 8080 and the domain, and deploy. From then on every push to the branch you chose ships a new version. Set a memory limit on the container so the heap follows it. See how much RAM a Java app needs.
Deploy with Claude Code or OpenCode
Coolify has a REST API. In the dashboard, open Keys & Tokens and create an API token. Give your coding agent the token and the server URL and ask it to create the application from your repository and deploy it. The agent uses the same API you would use from a script. You can limit what the token may do, and revoke it at any time.
Flat price vs. usage-based pricing
Example: a Spring Boot service with 4 vCPU and 8 GB of RAM running all month and sending 1 TB of traffic.
| Railway (usage rates) | Business, self-managed | |
|---|---|---|
| CPU | 4 × $20 = $80 | included |
| Memory | 8 × $10 = $80 | included |
| Traffic | 1,000 GB × $0.05 = $50 | 4 TB included |
| Per month | about $210 | €45.05 |
Railway rates as published on railway.com in November 2026: $20 per vCPU per month, $10 per GB of RAM per month and $0.05 per GB of egress, billed by usage. An idle app costs less there and a busy one more. Our price is the same every month. Our vCPUs are virtual cores on shared AMD EPYC hosts, and every plan is a separate server.
What it is not
You get one server, not a cluster that scales by itself. When your services outgrow it, move to a larger plan or add a second server to the same Coolify dashboard. For most Java teams, agencies and side projects, one well-sized server is plenty.
Questions about the app platform
What is Coolify?
Coolify is an open-source, self-hosted platform for deploying applications. It adds a web dashboard to your server where you connect a Git repository, choose how to build it and deploy. It also runs databases and issues Let's Encrypt certificates automatically.
Which size do I need for Coolify and a Java app?
Coolify itself needs about 2 GB of RAM and 2 CPU cores. Standard (4 GB) runs one small Java service. Business (8 GB) is the recommended size for a Spring Boot or Quarkus app with PostgreSQL, and Pro (16 GB) for several services.
Can Claude Code or OpenCode deploy for me?
Yes. Coolify has a REST API. Create an API token in the dashboard, give it to your coding agent with the server URL, and ask it to create and deploy the application. You can revoke the token at any time.
Is traffic billed per GB?
No. Each plan includes a monthly traffic allowance in its price: 3 TB on Standard, 4 TB on Business, 5 TB on Pro and 10 TB on Scale. There is no per-GB bill.
What is different from a cloud PaaS?
You get one server of your own, not an automatically scaling cluster. When you outgrow it, you move to a larger plan or add a second server to the same Coolify dashboard.
Who updates Coolify?
On Self-managed plans you do, with root access. On Managed plans we apply OS, Docker and Coolify updates, run daily offsite backups and monitor the server.