Skip to content
BSWEN
Blog

Free Hosting for Spring Boot: Best Options in 2026

Cowrie
Cowrie
Dev @ Bswen

Purpose

I built a Spring Boot application and wanted to deploy it without spending money. I found several free hosting options, but each has limitations that matter for Java applications. This post covers the platforms I tested, what worked, what didn’t, and what you should know before deploying.

Spring Boot needs memory. The JVM alone consumes 200-300MB before your application code runs. Most free tiers offer 512MB-1GB RAM, which is tight but workable if you tune the JVM. I’ll show you how.

AWS Free Tier (12 Months)

AWS offers 750 hours/month of EC2 t2.micro or t3.micro instances with 1 vCPU and 1GB RAM. This lasts 12 months, not forever. The 1GB RAM works for small Spring Boot applications, and you get full Linux control.

Pros:

  • 12-month free trial gives you time to test
  • 1GB RAM sufficient for small Spring Boot apps
  • Full Linux control (Ubuntu/Amazon Linux)
  • Professional cloud infrastructure

Cons:

  • Credit card required
  • Expires after 12 months
  • Setup complexity for beginners
  • Auto-charge if you exceed limits

Deployment: Build your executable JAR with mvn package. Launch an EC2 t2.micro instance (Ubuntu), allow port 8080 in security groups, SSH in, install Java 17+, and run:

Terminal window
java -jar app.jar

Or use systemd for a service. AWS Elastic Beanstalk is easier but adds complexity.

I tested this with a basic Spring Boot REST API. It ran fine with default JVM settings. Response times were good, and the app stayed running 24/7 without cold starts.

Best for: Production-ready apps, learning AWS, serious projects

Google Cloud Free Tier

Google Cloud offers an e2-micro instance (0.25 vCPU, 1GB RAM) in three US regions. Unlike AWS, this free tier never expires. You also get Cloud Run free tier: 2 million requests per month.

Pros:

  • Free tier never expires
  • 1GB RAM works for small Spring Boot apps
  • Cloud Run free tier (2 million requests/month)
  • Google’s infrastructure

Cons:

  • Limited to 3 US regions
  • 0.25 vCPU can be slow for Java apps
  • Cold starts on Cloud Run with JVM
  • Setup complexity

Deployment Options:

  1. Compute Engine: Deploy JAR directly to e2-micro instance
  2. Cloud Run: Containerize Spring Boot app (better for scale)

For Cloud Run, build a Docker image and deploy:

Terminal window
gcloud run deploy --image gcr.io/PROJECT_ID/app --platform managed

I found Cloud Run cold starts took 10-20 seconds for Spring Boot. The e2-micro VM was slower than AWS due to 0.25 vCPU, but the never-expiring free tier is compelling for long-term projects.

Best for: Long-term projects, Google ecosystem users, containerized deployments

Azure Free Account

Azure offers 750 hours/month of B1s burstable VM (1 vCPU, 1GB RAM) for 12 months, plus $200 credit for the first 30 days. Azure App Service also has a free tier with limitations.

Pros:

  • $200 credit flexibility
  • App Service for easier deployment
  • Good GitHub integration
  • 12 months free compute

Cons:

  • Credit card required
  • App Service Free Tier has limitations (custom domain, SSL)
  • Complexity for beginners
  • 12-month expiration

Deployment:

Azure App Service is the easiest option. Build your JAR with Maven, create a Web App resource, and deploy via GitHub Actions or Azure CLI. Azure handles the Java runtime automatically.

I tested App Service Free Tier. Deployment was straightforward, but the free tier limitations annoyed me - no custom domain and SSL restrictions. The B1s VM worked better.

Best for: Enterprise developers, Microsoft ecosystem, App Service simplicity

Oracle Cloud Always Free

Oracle Cloud offers 2 AMD-based VMs (1 ocpu, 1GB RAM) and 4 ARM-based Ampere A1 cores (24GB RAM total) that are always free. This never expires and is the most generous free tier I found.

Pros:

  • Truly free forever (no trial expiration)
  • Generous ARM resources (24GB RAM!)
  • 1GB AMD VM sufficient for small Spring Boot apps
  • Professional infrastructure

Cons:

  • Oracle console is complex
  • ARM requires architecture-specific builds
  • Setup complexity for beginners
  • Documentation can be confusing

Deployment:

Use the AMD VM (1 ocpu, 1GB RAM) for simplicity. Install Java 17+, deploy your JAR or use Docker, and set up firewall rules. The ARM VMs offer more resources but require building for ARM architecture.

I found the Oracle console frustrating compared to AWS or Google Cloud. But once configured, the free tier is unmatched. I recommend this for advanced users who want free forever hosting.

Best for: Advanced users, long-term projects, free forever hosting

Railway and Render

Railway and Render are newer platforms with free tiers focused on developer experience.

Railway offers $5 free credit monthly with 512MB RAM and 0.5 vCPU. The credit renews each month but runs out if you use it. Automatic deployments from GitHub, built-in PostgreSQL and Redis, and good logs.

I found Railway’s UI beautiful and intuitive. GitHub integration worked flawlessly. But the $5 credit is a moving target - heavy usage means you’ll pay or wait for reset.

Render offers 512MB RAM and 0.1 CPU with automatic HTTPS and GitHub integration. The free tier spins down after 15 minutes inactivity. Cold starts took 30+ seconds for Spring Boot.

I tested Render with a simple Spring Boot app. The JVM needed tuning to fit in 512MB:

Terminal window
JAVA_OPTS="-Xmx400m -Xss256k -XX:+UseContainerSupport"

Even with tuning, the app crashed with OutOfMemoryError under load. But for light prototypes, the zero-config deployment is appealing.

Best for: Quick prototypes, GitHub integration, beginners

Fly.io

Fly.io offers 3 small VMs with 256MB RAM and shared CPU, always free. Global deployment is the selling point.

Pros:

  • Truly free (not trial)
  • Deploy close to users worldwide
  • Docker-based deployment
  • Good CLI tool

Cons:

  • 256MB RAM is too small for most Spring Boot apps
  • Would need to upgrade to paid tier
  • More complex than Render/Railway

I tested Fly.io with a minimal Spring Boot app. It wouldn’t even start with default settings. After tuning to 200MB heap, the app started but crashed under first request. I don’t recommend Fly.io for Spring Boot unless you pay for more RAM.

Best for: Apps that can scale globally, Go/Rust apps (not Java)

Free Tier Limitations

Free tiers work, but you must understand their limitations.

Memory constraints are the biggest issue. Spring Boot needs minimum 512MB RAM, but 1GB is recommended. The JVM needs 200-300MB overhead, plus application memory.

Cold starts affect serverless and container platforms. Render, Railway, and Fly.io spin down after inactivity. Spring Boot cold starts take 10-30 seconds depending on app size.

Resource limits include CPU throttling, network bandwidth caps, and storage quotas. AWS and Google Cloud enforce these strictly.

Expiration matters. AWS and Azure free tiers expire after 12 months. Google Cloud, Oracle Cloud, and some PaaS platforms offer permanent free tiers.

JVM tuning is mandatory for 512MB tiers:

Terminal window
# For 512MB RAM (Render, Railway)
java -Xmx350m -Xss256k -XX:+UseContainerSupport -jar app.jar


# For 1GB RAM (AWS, GCP, Azure)
java -Xmx700m -Xss512k -XX:+UseContainerSupport -jar app.jar

Comparison Table

PlatformFree RAMCPUTime LimitCold StartsBest For
AWS (12mo)1GB1 vCPU12 monthsNoProduction, learning AWS
GCP (Always)1GB0.25 vCPUNeverNo (VM), Yes (Cloud Run)Long-term projects
Azure (12mo)1GB1 vCPU12 monthsNoEnterprise, App Service
Render512MB0.1 vCPUForeverYes (15min)Beginners, GitHub
Railway512MB0.5 vCPU$5/mo creditYesDX, databases
Fly.io256MBSharedForeverYesGlobal apps (not Java)
Oracle1GB-24GB1-4 vCPUNeverNoAdvanced, free forever

When to Upgrade

You’ll know it’s time to pay when:

  • App crashes with OutOfMemoryError
  • Users complain about slow response times
  • You need background workers or scheduled jobs
  • Custom domain and SSL are required
  • Production reliability matters

Cost comparison after free tier:

  • AWS EC2 t3.medium: ~$15/month (2 vCPU, 4GB RAM)
  • GCP e2-small: ~$10/month (2 vCPU, 2GB RAM)
  • Render Pro: $7/month (1GB RAM)
  • Railway: $5-20/month (512MB-4GB RAM)
  • DigitalOcean: $6-12/month (1-2GB RAM)

Summary

After testing these platforms, here’s what I recommend:

Best for beginners: Render. GitHub integration, zero configuration, automatic SSL. The 512MB is tight, but for learning and prototypes it works.

Best for long-term free: Oracle Cloud Always Free. The 24GB ARM resources are unmatched, but the learning curve is steep. Skip unless you’re comfortable with Linux and networking.

Best for production: AWS Free Tier. 12 months gives you time to prove your concept. 1GB RAM is sufficient for small apps, and you learn AWS in the process.

Best for containerization: Google Cloud Run. 2 million free requests per month, good auto-scaling. Cold starts are the downside.

Key takeaways:

  1. Free tiers have memory limits - tune JVM accordingly
  2. Spring Boot needs 512MB-1GB minimum
  3. Container deployment is more flexible
  4. Most free tiers spin down (cold starts)
  5. Plan your migration path before free tier expires

Build your Spring Boot executable JAR, choose a platform based on your needs, deploy and monitor memory usage, optimize JVM settings, and plan your upgrade path before the free tier expires.

Final Words + More Resources

My intention with this article was to help others share my knowledge and experience. If you want to contact me, you can contact by email: Email me

Here are also the most important links from this article along with some further resources that will help you in this scope:

Oh, and if you found these resources useful, don’t forget to support me by starring the repo on GitHub!

Comments

Tags: