I recently discovered a new PaaS platform, Cloady (operated by NeptoLab), similar to Railway / Render. Its free tier is quite generous, so I’m sharing it with everyone.
-
Free tier
• Per workspace: 0.5 vCPU, 1 GB RAM, 5 GB SSD
• Registration only requires an email address; log in with an email verification code, with no card required at any point
• The free plan has a “workspace-wide limit”: as long as all applications combined stay within the limits above, the cost is $0/month; you only need to add a card and pay once you exceed them
• Each web application is automatically assigned an xxx.cloady.io subdomain with a built-in TLS certificate -
Available regions (free tier)
• EU1 Germany
• USE1 US East
• USW1 US West
Note: The JPN1 Japan region is currently not accepting new free applications. New workspaces may default to JPN1, so remember to switch regions manually when deploying. Singapore, the UK, India, and other regions are not available yet. -
Supported deployment methods
• GitHub repositories (automatic deployment on push)
• Docker images
• Direct project-folder uploads (automatically builds if a Dockerfile is present)
• Template marketplace: one-click installation of common applications and databases
• Import from Railway / Render / Vercel
• REST API, CLI, Terraform, and MCP are also supported, allowing AI clients to deploy directly on your behalf -
Getting started
-
Go to cloady.com, register with an email address you regularly use, and log in with the six-digit verification code
-
The system will automatically create a workspace
-
Click Deploy and select a source (GitHub / image / upload / template)
-
Select EU1, USE1, or USW1 as the region, keep resource usage within the free tier, and confirm that the page shows $0.00/mo
-
Once deployment is complete, you can access the application using the cloady.io address shown on its page
-
Tips to avoid pitfalls
• The web interface’s “Docker image” deployment does not let you set the port or startup command. After creation, the service, port, and environment cannot be changed either. If the image’s default port is not detected, the application will be “inaccessible from the public internet.”
• Solution: write a simple Dockerfile (FROM the image you want + EXPOSE the port) and deploy via “Upload” or GitHub; alternatively, use the API’s deploy endpoint, which accepts port and startCmd parameters directly.
• The free tier is shared across the entire workspace. Opening several applications may exceed the limit. Without a linked card, you will simply be prompted to pay; you will not be charged unexpectedly. -
Paid pricing (for reference)
• Small: 1 core / 2 GB / 50 GB, $5/month
• Medium: 4 cores / 8 GB / 200 GB, $15/month
• Large: 10 cores / 44 GB / 600 GB, $65/month
• Max: 16 cores / 92 GB / 1 TB, $120/month
Billing is based on the stated CPU, memory, and disk allocations; intermediate configurations are calculated proportionally. -
Terms reminder (please follow them carefully)
• Use a real email address that can receive messages normally; do not use a temporary email address
• Registering multiple free accounts in bulk is prohibited
• Mining, Tor, IRC, and similar uses are prohibited
• The new terms take effect for new accounts from 2026-11-02 and for existing accounts from 2026-11-16
Please take advantage of the free tier responsibly—don’t overexploit the platform and ruin this limited benefit for everyone.
What is it suitable for: personal blogs, bots, small APIs, monitoring probes, and test environments. With 0.5 cores and 1 GB of RAM, it is enough to run most lightweight services.
The information above is current as of 2026-11-03 and is based on my own testing. Platform policies may change, so refer to the official website.