AgentSpherePublic apps stay public
Customer-facing apps are securely accessible over HTTPS with proper domain setup, SSL and a hardened configuration.
SERVER SETUP BLUEPRINT
The server setup that lets you sleep.
Secure public and private apps on your own VM. Recovery proven. Executed by your AI agent in about 45 minutes.
- 100+servers tested
- 10+ yearsexpertise
- German GmbHEssen
- 30-daymoney-back guarantee
Public appsHTTPS access
Secure VM
Private toolsVPN only
Encrypted backups
Verified on
Google Cloud
Azure
Hetzner
Hostinger
DigitalOcean
Contabo
Private tools stay private
Dashboards, CRMs, internal tools and databases are only reachable through a WireGuard VPN, kept off the public internet.
Recovery you can trust
Encrypted off-server backups and an actual recovery test on a fresh VM. No more hoping your backups work. You will know they do.
How it works
A tested path from empty server to verified recovery in 3 steps.
- 1
Give your AI agent the blueprint
Hand over the blueprint and a few basics about your project and environment. No server has to exist yet.
- 2
Let it configure your VM
Your AI agent sets up the server, installs and configures everything automatically.
- 3
Verify recovery on a fresh server
Follow the included test to restore from backup on a new VM and verify everything works.
What builders say
Real feedback from founders, self hosters and technical teams who ran the blueprint on their own servers.
4.8/5from 15 reviews
I can build features with Cursor all day, but deploying them securely is a different story. My AI agent followed the blueprint and got the server, HTTPS and private access configured in under an hour. Having n8n accessible only through the VPN is exactly what I wanted. The best moment was watching it restore everything on a fresh VM. That made the whole setup feel much more real.
- Apps
- Next.js SaaS, PostgreSQL, private n8n automation
- VM provider
- Hetzner
I was uncomfortable putting our customer-facing website and internal business tools on the same server. I knew I needed them separated but had no idea how to do it. I still had to connect to the VPN and provide some account details, but the instructions were understandable. Seeing the private admin panel disappear when I disconnected the VPN was especially reassuring.
- Apps
- Customer booking platform, private CRM, internal admin panel
- VM provider
- Hostinger
I could have configured the reverse proxy, VPN, backups and access rules myself. But that’s easily two or three days once you include researching the right architecture and checking everything. This gave my coding agent a tested starting point. I still inspected the configurations, but I wasn’t designing the infrastructure from scratch. That’s a great trade-off for €59.
- Apps
- FastAPI backend, React frontend, private Metabase analytics
- VM provider
- AWS EC2
We had a working prototype, but I wasn’t comfortable going live without knowing how we’d recover it. The backup configuration was useful, but the recovery test was the real selling point. It actually went through the restore on a fresh machine and checked the result. That’s much more convincing than a tutorial telling you your backups are probably fine.
- Apps
- Next.js customer portal, REST API, private operations dashboard
- VM provider
- Google Cloud
The result was great, but it wasn’t quite as hands-off as I expected. I needed to install some tooling on my laptop, sort out domain records and install WireGuard. None of it was particularly difficult with the instructions, but it took longer than expected. Once it was working, I loved having a setup where my client portal was public and everything administrative stayed private.
- Apps
- Client portal, Directus CMS, private admin dashboard
- VM provider
- DigitalOcean
I wanted to run several small services without paying separately for every managed platform. Every tutorial I found had a different opinion about reverse proxies, networking and backups. This blueprint gave the agent one consistent approach. The setup took a little longer than 45 minutes including account preparation, but I saved far more time than that by avoiding the usual research rabbit hole.
- Apps
- Node.js SaaS, background worker, private n8n workflows
- VM provider
- Contabo
We build automations for clients, so we needed public webhook endpoints without exposing our internal workflows and management tools. This separation was exactly what we were looking for. Our AI coding agent handled most of the configuration, and we ended up with something we can use as the starting point for future projects. The backup and recovery testing was a huge bonus.
- Apps
- n8n, client webhook endpoints, internal CRM, monitoring dashboard
- VM provider
- Hetzner
The blueprint made the server configuration much easier, but preparing my cloud permissions and access credentials took longer than I expected. I also needed to check a few things with the AI agent along the way. Once those prerequisites were sorted out, the networking and backup steps were straightforward. The final restore test gave me confidence that the work was actually complete.
- Apps
- SaaS beta application, private administrative dashboard
- VM provider
- Microsoft Azure
I’ve deployed plenty of servers, but properly testing disaster recovery is always the part that gets postponed. This workflow made it part of the initial setup rather than another item on my backlog. The actual restore exercise was the most valuable part for me. I could inspect what happened, verify access and feel much better about putting something important on the machine.
- Apps
- Node.js API, PostgreSQL, private Grafana dashboard
- VM provider
- AWS EC2
My app was ready, but I had been avoiding the hosting decision because every option seemed complicated. I didn’t want to become a systems administrator just to launch a business. The blueprint made the process feel structured. I answered questions, gave my agent the required information and followed the checkpoints. I still have things to learn about operating a server, but getting started felt manageable.
- Apps
- Subscription web app, private customer-support dashboard
- VM provider
- Hostinger
I had three small applications that didn’t justify separate servers. I wanted a clean way to host them together without accidentally exposing the monitoring dashboard. The blueprint’s public/private separation made sense immediately. What I liked most was having one defined configuration and recovery process instead of a collection of commands from old projects.
- Apps
- Two client-facing apps, staging site, private Uptime Kuma monitoring
- VM provider
- DigitalOcean
Our MVP went from idea to working product incredibly quickly because of AI coding tools. Infrastructure was the first part where that speed broke down. Giving the agent a tested blueprint rather than asking it to invent a server setup was a much better experience. I particularly liked having explicit checkpoints and knowing there was a recovery procedure if something went wrong.
- Apps
- Python API, React frontend, background processing service, private operations UI
- VM provider
- Google Cloud
Very solid starting point and definitely faster than building everything from scratch. I reviewed the networking configuration and changed a few defaults for my environment. The backup setup was easy enough, although the fresh-VM recovery test took some extra time while I checked the restored services. I’d love an even clearer preflight checklist and a single summary of all the generated configuration.
- Apps
- Next.js frontend, FastAPI backend, PostgreSQL, internal admin UI
- VM provider
- Hetzner
We started with multiple services on different platforms and spent more time managing deployments than we wanted. Our goal was to simplify the infrastructure for an early-stage product. The setup gave us a structured way to host our application services and keep internal tools private. The recovery test was important because we didn’t want to reduce complexity at the cost of reliability.
- Apps
- B2B portal, private billing operations UI, internal reporting tool
- VM provider
- Microsoft Azure
I’ve done this setup several times and always end up searching through old notes, scripts and half-remembered commands. Having a repeatable blueprint that my coding agent can follow is worth the price alone. It removes the boring setup work while still letting me review the result. And unlike my old scripts, this approach actually includes testing recovery.
- Apps
- Two micro-SaaS projects, private n8n instance, internal analytics
- VM provider
- Contabo
See if it fits youA few questions to find out if it is the right fit.
AFTER YOU BUY
You paste one prompt. Your agent does the rest.
Checkout sends you straight to your access page. It carries a ready-made prompt with your private download link already in it. Copy that into Claude Code or Codex and the run begins.
- Pay and land on your access page
- Copy the prompt we generated for you
- Paste it into your AI agent and approve each step
Your start prompt
Download my AI VM Setup Blueprint: <your private download link> Unpack it, read START-HERE.md, then ask me whether to start the blueprint now.
Shown here without a link. The real prompt is generated for you after payment and is personal to your order.
FIND YOUR STARTING POINT
How exposed is your self-hosted setup?
Three or four short questions show where the blueprint helps and whether the agent-led path fits the way you use AI.
What is your current situation?
WHAT YOU GET
Everything the agent needs, and everything it leaves behind.
One versioned package your AI agent reads directly, and a documented result you keep.
The blueprint package
A versioned zip with stage files, commands, checks and acceptance criteria your agent reads directly.
A hardened VM
Pinned supported image, restricted host access, a firewall and a documented recovery path.
Separated public and private routes
Customer traffic and internal administration never share the same entrance.
Encrypted off-server backups
Scheduled, encrypted and stored away from the VM they protect.
A recovery test you actually run
Restore on a fresh VM, with the result and measured recovery time written down.
A handover you own
Sanitised configuration, evidence and day-two operating notes stay with you — no lock-in.
WHO THIS IS FOR
Built for people who ship, not for people who tune servers.
You decide and approve. Your agent does the repeatable infrastructure work.
Founders shipping a product
You need customer-facing apps online and internal tools private, without hiring an ops person first.
Self hosters and small teams
You already run a VM. You are not sure you could rebuild it if the disk disappeared tonight.
Technical teams without a platform engineer
You know what good looks like. You want a tested path instead of assembling one from twelve blog posts.
Probably not for you if
- You need high availability or zero-downtime failover — this is a single-node setup.
- You want someone else to run your infrastructure for you.
- You have never used an LLM and do not want to start with a CLI agent.
WHY THIS EXISTS
Getting online is the easy part.
Installers, snapshots and backup jobs each solve a piece. Nothing connects them and proves the way back.
An installer
Gets software or a cluster running, then stops.
A backup job
Copies data, but never proves the whole recovery path.
This blueprint
Builds the base, restricts access, verifies protected backups and records what was actually restored.
Single-node recovery is not high availability or zero downtime.
A server of your own. Explained your way.
Same blueprint. Choose the level of detail that suits you.
AI VM SETUP BLUEPRINT · FOR NON TECHIES
What is a VM? A home for your digital tools.
A virtual machine (VM) is a computer you rent online. It stays available when your laptop is closed: a place to host the website AI created for you, your own CRM system or an automation tool like n8n. Several compatible services can share one VM, as long as it has enough capacity.
Your server · Your accounts · A foundation for multiple services
FROM AN IDEA TO A SERVER
You don’t need to know every command.
The blueprint gives your AI agent a guided plan to prepare that computer: access, security, backups and checks. You get explanations in smaller steps, with documented defaults for unfamiliar technical choices.
Tell it what you need
Share what you want to host, your experience and your budget. The agent uses this to plan the server size and setup.
Review the plan
You provide your accounts and approve the named resources and costs. Your agent explains the changes before carrying them out.
Get a tested foundation
The agent prepares the server, checks access and backups, and leaves instructions for running it. Deploying your apps comes next.
What you bring
- A cloud account and a domain you control.
- An AI coding agent on your computer, with a suitable plan or API budget.
- A backup destination and a private Git repository; the agent can guide their setup.
- Time to review decisions and supervise changes. Hosting, storage and AI usage cost extra.
The server is the starting point.
This is a downloadable blueprint that your AI agent executes. It does not include hosting or ongoing server management. App deployment and migration are separate next steps, with guides included. You remain responsible for operating the server; one VM does not provide uninterrupted availability.
Your time. Your peace of mind.
Spend less time fixing. Feel ready to recover.
You already need an AI agent to do the work. The blueprint gives it a sequence, checks and a clear finish, so you can spend more time using your tools.
The expectation
“My agent will sort it in half a day.”
Getting a website online can look quick. Making several tools work together, keeping private areas private and checking that your data can come back takes more work.
Our own setup experience
Five days later, still chasing issues.
What looked like half a day became around five days: roughly 10× the planned time, with networking and backups still leaving unanswered questions. This is our experience, rather than a promised timeline for your setup.
A clearer route with the blueprint
Know what is finished. Know how to get it back.
- Your agent follows a defined order instead of reinventing the setup.
- Checks catch problems before you put your tools to work.
- Backup material is restored in isolation and the result is recorded.
- A handover shows what passed and what still needs attention.
The good feeling comes from evidence: you have seen a restore work, and you have a record to return to when something goes wrong.
Want proof that the whole server can be replaced? Choose the optional fresh-server recovery rehearsal. Full recovery is marked proven only after that rehearsal passes.
WHY THE SETUP MATTERS
Getting online is the start. Keeping control is the work.
When several tools share a server, a small mistake can affect more than one app. Here is what the blueprint helps you prepare for.
Why separate your services?
What can go wrong
Your website should not automatically have access to the data in your CRM. Without deliberate boundaries, a compromised app can put other tools and their data at risk.
How the blueprint helps
The blueprint prepares the platform’s base security and isolation. Its follow-up workload guide covers separate identities, secrets and network rules for each app. Apps still share one machine; these boundaries are not complete isolation.
Why have different ways in?
What can go wrong
Visitors need to reach your website. They do not need to reach your server controls or private tools. If everything uses the same public entrance, it is easier to expose something you meant to keep private.
How the blueprint helps
Separate public and private entrances — called ingress — give each kind of traffic its own route. Public sites use HTTPS; administration and internal routes use private access.
Why keep backups somewhere else?
What can go wrong
A server can fail, a disk can be lost or data can be deleted by mistake. A backup stored only on that same server can disappear with the original.
How the blueprint helps
The blueprint sets up protected, encrypted backups outside the VM. When you add apps, their data and backup coverage must be included through the workload guide.
Why test that a backup restores?
What can go wrong
A green “backup complete” message does not tell you whether the copy is usable. Missing files or a missing decryption key may only become obvious when you need them most.
How the blueprint helps
The backup stage decrypts and restores backup material in isolation and records the result. That gives you evidence of a restore, with any remaining gaps documented.
Why prove full recovery?
What can go wrong
After losing a server, you need more than data: settings, access, secrets and instructions must come together on a replacement. Discovering missing pieces during an outage costs time.
How the blueprint helps
An optional full recovery rehearsal rebuilds the platform on a fresh VM and records what worked and how long it took. “Recovery proven” applies only when that rehearsal was actually run and passed. It does not mean zero downtime.
Why keep a handover?
What can go wrong
Months later, you may not remember what the AI changed. Without saved configuration and instructions, an update or repair can turn into guesswork.
How the blueprint helps
You keep a private configuration repository, operating instructions and a record of checks and open issues. Status endpoints can connect to an uptime service you own; ongoing monitoring and maintenance remain your responsibility.
AI VM Setup Blueprint
Build the recovery path before you need it.
Get the current complete, agent-executable Blueprint as a zip download. One company may use it across unlimited internal and client VMs.
Please try again later or contact us.PREFER A HAND?
Guided Launch · €999
Your AI performs the implementation. We review the architecture, help resolve important decisions and join you for the final acceptance and recovery session. Includes a kickoff, limited asynchronous support and a final evidence review; migration is separate.
A few useful answers.
Can I use this without being a server expert?
The guided mode explains changes at your experience level and uses documented technical defaults. You still need to use an AI coding agent, provide accounts, review costs and supervise the work. If you want someone to operate the server for you, ask about guided help.
Will my website and tools be installed?
This blueprint prepares the server foundation. It includes next-step guides for deploying or migrating workloads, but the initial setup does not install your applications.
What am I paying for?
The blueprint package and its workflow. Cloud hosting, your domain, backup storage and AI usage are separate costs paid through your own accounts.
What happens if the server fails?
The setup includes protected off-server backups and restore verification. You can also select a full recovery rehearsal on a fresh VM. Recovery takes time; a single server is not a high-availability system.