AI-native cloud

The European cloud your AI agent runs end to end

Your agent provisions the VM and a managed database, deploys your app, and keeps it running. You set up nothing. Point it at a project through the redu MCP and it does the whole thing on real EU infrastructure. Works with any MCP-capable agent. Claude Code is the example here.

Start buildingRead the deploy guide
Install the MCPclaude mcp add --transport http redu-cloud https://mcp.redu.cloud/mcp

Then tell your agent: "Deploy this repo to redu, in the EU."

What your agent does on redu

  • Provisions a virtual machine, a managed PostgreSQL or Redis, private networks, and autoscaling clusters.
  • Deploys a Git repo or a docker-compose project to a live https URL, with DNS and TLS wired up.
  • Keeps it running, and when a deploy breaks it fixes it in place and pushes the fix back to your repo.
  • All from one prompt, so you set up nothing yourself.

On other clouds, you assemble it yourself

  • You provision the server, configure it, and set up TLS before anything runs.
  • You set up and operate the database yourself, or wire in a separate managed service and a separate bill.
  • Your agent can connect to a box you already built, but the setup was on you.
  • Fine when you want to run the plumbing. redu is for when you would rather your agent did.
How it compares

redu vs a bare VPS vs a PaaS

A bare VPS gives your agent a shell but leaves the setup to you. A PaaS hides the machine so your agent cannot operate it. redu gives your agent the whole thing: it provisions, deploys, and runs real EU infrastructure for you.

Capability
Bare VPS or PaaS
redu
Who sets it up
Bare VPS: you do it (server, TLS, database). PaaS: the platform, within its limits.
Your agent, end to end, from one prompt.
Managed Postgres and Redis
Bare VPS: you run it yourself. PaaS: yes, separate and usually premium.
Provisioned and auto-wired to the app.
Deploys your app
Bare VPS: you script it. PaaS: git push.
Point your agent at the repo through the redu MCP.
Keeps it running and fixes it
Bare VPS: you do it by hand. PaaS: the machine is hidden, so nobody can.
Your agent operates the real machine and fixes a broken deploy in place.
Control of the machine
Bare VPS: full, once you built it. PaaS: none, there is no shell.
Full. Real infrastructure your agent can SSH into and operate, not a sandbox.
Location and billing
Depends on the provider, and a PaaS is usually US regions and per-unit pricing.
EU data residency in Germany, billed by the hour.
Why teams pick redu

A real EU cloud your agent runs, not a box you set up or a sandbox it cannot touch.

Your agent does the setup, deploys the app, and keeps it running, on real infrastructure you fully control, in the EU.

Your agent does the setup, not you

From one prompt, your agent provisions the VM and a managed database, deploys your app, and wires DNS and TLS. On other clouds that setup is on you before anything runs.

It deploys and it heals

Deploys recreate until the app is actually running. When something breaks later, your agent fixes the live machine in place and pushes the fix back to your repo.

Managed data, provisioned and wired

Managed PostgreSQL and Redis come up with the app and the connection string is injected. DATABASE_* aliases are set so Rails, Django, and Prisma work unmapped.

Real infrastructure, not a sandbox

Your agent operates real VMs with full control, including SSH, so it can actually run and fix your app. EU data residency, billed by the hour.

The honest trade

A fully-managed PaaS git push is frictionless, and redu does not have compliance certifications like HIPAA or SOC 2 yet. If you need those today, redu is not your move yet. If you want a real EU cloud your agent runs end to end, this is it.

FAQ

AI agent cloud questions

Practical answers for running your infrastructure from your AI agent.

Can an AI agent set up and run infrastructure on redu?

Yes, end to end. Over the redu MCP, any MCP-capable agent (Claude Code is a common example) provisions VMs and managed databases, deploys your app, and operates the running machine.

Do I have to configure the server myself?

No. Your agent handles provisioning, deployment, DNS, and TLS from one prompt. You can also drive redu yourself like a normal cloud if you prefer.

Can my agent SSH into the machine?

Yes. These are real virtual machines, not a sandbox, so your agent has full control to run commands, read logs, and fix issues in place.

Is redu in the EU?

Yes. redu offers EU data residency, with servers in a German data centre. Redu Cloud Ltd is a UK company registered in England and Wales.

Does it include a managed database?

Yes. Managed PostgreSQL and Redis are provisioned alongside your app and auto-wired, with the connection string injected. DATABASE_* aliases are set so Rails, Django, and Prisma work unmapped.

What does it cost?

Everything is billed by the hour. Most business apps run on a VM from about £20 to £35 a month, with no per-unit tax and no surprise bill after a busy weekend.

Which agents work with redu?

Any MCP client. Claude Code is the common example, but redu is agent-agnostic.

Keep exploring

See how redu works with your agent.

AI and MCPPricingCompareDeploy guide
Start today

Let your agent run it. Get a live URL.

Install the redu MCP, tell your agent to deploy your repo, and it provisions, deploys, and keeps your app running on real EU infrastructure. New accounts start with £200 in credits.

Start building