, it’s been anything but calm this week. Perfect timing, then, to roll out something new. Amid the chaos, we’ve been quietly putting together the first version of the 

. 
We teased it last week, and now it’s here. Yes, yes we know it can feel like — 

“just one more MCP server, bro. I swear this one’s different.”

 But in all honesty, we think you’ll like what it can do. More on that below.

. We’ve got the full breakdown further down — and trust us, there’s some seriously cool stuff in there.

Now, let’s dive into what shipped this week!

Databases are first-class citizens on Railway, and we want them to deliver the same 

oh-my-god-I-can’t-believe-it-was-this-easy

 experience you get everywhere else on the platform. The zero-to-one experience already delivers: you can 

Now we’re shifting our focus to what happens 

. We're hardening databases on Railway so they’re more reliable and scalable out-of-the-box. You’ll get more opinionated workflows while needing less database-specific knowledge, letting you focus on shipping your apps. Our goal is that you never need to look beyond Railway when building your app, database included. 

We’ve been shipping improvements to make managing

just as seamless as creating them. Here’s a quick look at the latest updates for databases and volumes — and a preview of what’s next.

You can now pick a mount path from a volume's history. This way you don't have to type it out and risk making a mistake

We improved the volume name generator to make volume names more clear like 

We made the restore backups button a lot less scary

You can try out all of these updates by going to 

Automatic minor/patch version upgrades for your database

your feedback on your experience with databases on Railway. You can drop your suggestions and issues 

Although it's still early and highly experimental, it's ready for you to start using. 

You can have your coding agent spin up a project, attach a domain, search and deploy a template, or pull logs from a running service without leaving your editor. 

By design the MCP server doesn’t include any destructive actions, so you shouldn’t worry about your project resources suddenly disappearing. We also strongly recommend that you try out the MCP server in an isolated environment.

This is the initial release. We’re curious which workflows and features you’d want next, and what’s missing from your day-to-day MCP setup. Share your thoughts in 

In the coming weeks, we’ll do a deep dive into how it’s built and share some behind-the-scenes (including a few moments where it felt like things wouldn’t work at all).

 is officially in the books — and wow, did the community deliver. There were two categories: 

 which is all about ambitious builds, and 

, which was all about teaching and sharing.

Switchyard, an autoscaling infrastructure platform

Pushflow, a web push notification management system

Remotion, a video generation server guide

Railway + Terminus, an e-ink dashboard for deployments

 came out of the hackathon. You can explore the community’s favorites in the 

 — and maybe start thinking about what you’ll build for the next one.

We improved the shared variables experience by enabling you to copy a configuration from one shared variable to others

We improved the Railway GitHub PR bot to support multiple projects linking to the same PR. Previously, each project would overwrite the last project's PR comment update, but now the bot will happily merge updates and show all of them in one comment. 

This significantly improves the experience of working with monorepos linked to multiple projects

 flag, allowing you to create a project non-interactively from the CLI, which is especially useful in CI environments

https://res.cloudinary.com/railway/image/upload/v1755279419/2025-05-15_pjuxa0.png

Better Databases, Railway MCP Server, Hackathon winners

Changelog #0251