The box that grows the business.
Your growth stack on hardware you own. Pick a plan and the box builds itself — Growth OS, n8n, Postgres, Redis, auto-TLS and daily backups on your own isolated server, live at https://you.growvps.com in about 5 minutes. We manage it. You own it.
The deal
Three steps. One of them is yours.
Pick a plan
Flat monthly for CPU boxes, metered for GPU. Checkout takes a minute.
The box builds itself
Zero human touch — provisioned, installed, secured and certified automatically.
Log in and grow
Your stack is waiting at your own subdomain, TLS included.
The fleet
Seven AI marketing agents. Preinstalled. On duty 24/7/365.
Growth OS isn't a dashboard you have to drive. It's an always-on system of specialized AI marketing agents — one per channel — coordinated by a central orchestrator and supervised by you, generating and converting leads around the clock. The whole fleet comes preinstalled on your box.
The center of the constellation — HELIOS coordinates every agent in the fleet, so the channels move as one operation instead of seven silos.
Every agent works supervised — humans stay in the loop, and nothing the fleet drafts goes out without approval. More on that below.
The command center
One login runs the whole operation
The fleet does the work; the Command Center is where you watch it, steer it and approve it. Every module below comes preinstalled on your box, already wired to the agents — nothing to integrate, nothing to subscribe to separately.
Agents draft. You approve.
Human-in-the-loop isn't a feature toggle — it's how the whole fleet operates. What the agents draft lands in Approvals, and nothing goes out until you say so. Watch the dot wait for the green check: that pause is you.
Leads
Every lead the fleet captures, in one place — fed around the clock by the five capture agents.
Pipeline
Leads become deals and move stage by stage toward closed — visible at a glance.
Approvals
Human-in-the-loop, built in: agents draft, you approve. Nothing ships without your sign-off.
Reactivation
Dormant-lead revival — the leads you already paid for get a second life.
SDR / Outreach
The outbound side of the house, run right next to the leads it touches — no separate tool.
Spend & Margin
Spend and margin tracking across the operation — what the fleet costs and what it brings back.
Activity feed
A live feed of everything the fleet does, as it happens.
Team Chat
Built-in team chat — your people and your pipeline in the same place.
Plus Dashboard, Clients and Fleet view
And underneath it all: a private n8n automation engine wired to the stack, with Postgres 16 and Redis 7 as your data layer. All of it arrives preinstalled on your box, before you first log in.
Replace your stack
Six subscriptions. Or one box.
Most growth teams are paying six subscriptions to rent a stack they could own: n8n cloud, managed Postgres, managed Redis, SSL, backups, and a VPS underneath it all. On your box, all of it is one flat price — on hardware you can export and walk away with.
| What you'd subscribe to | Typical entry price* | On your box |
|---|---|---|
| n8n cloud plan | ~$24/mo | ✓ included — your own private n8n |
| Managed Postgres | ~$15/mo | ✓ included — Postgres 16, yours to query |
| Managed Redis | ~$10/mo | ✓ included — Redis 7 |
| SSL / proxy service | $0–10/mo | ✓ included — auto-TLS, issued and renewed automatically |
| Offsite backup service | ~$5–10/mo | ✓ included — daily encrypted incremental snapshots |
| A VPS itself | ~$10–30/mo | ✓ it is the box |
| Total, separately | roughly $64–99+/mo, six bills | $49–99 flat. One bill. Exportable. |
*Approximate typical entry prices for comparable managed services. Your mix may differ; no affiliation with any provider implied. The point stands either way: it's fewer bills on hardware you own the data on.
What you get
Everything preinstalled, already configured
No testimonials, no logo wall. The proof is the architecture — this is exactly what comes up on your box, wired together before you ever log in.
Auto-TLS edge
TLS certificates are issued and renewed automatically at the edge. You never think about SSL again.
- Real certificate on https://you.growvps.com
- Renews itself, forever
- Routes / → Growth OS, /n8n/ → n8n
Growth OS
The full app on a server nobody else shares — not a seat in someone's multi-tenant pool.
- Isolated on your own box
- Environment pre-generated at build time
- Data on its own named volume
n8n
A private automation engine. Build workflows that shared platforms cap or forbid.
- Lives at /n8n/ on your subdomain
- Own Postgres database, per-box encryption key
- Webhook URL pre-configured to your subdomain
Postgres 16
A real database, yours to query, dump and export — not an API quota.
- Credentials generated per box
- Separate growthos and n8n databases
- Data on its own named volume
Redis 7
Cache and queue for the stack, ready on day one.
- On the stack's private internal network
- Data on its own named volume
Encrypted backups
Daily encrypted incremental snapshots to offsite object storage, in a standard open format — restorable anywhere, with or without us.
- App + database + automation volumes
- 7 daily + 4 weekly kept, pruned automatically
- Runs once offsite storage is configured for your box
The server itself
Hardened at build time by our zero-touch installer, before the stack even starts.
- SSH keys only — non-root sudo user, password login disabled permanently
- ufw firewall, default deny — only ports 22, 80 and 443 open
- fail2ban — brute-force attempts get banned automatically
- Unattended security upgrades — patches land without you asking
- Open-source containers — standard, portable, nothing proprietary locks your data in
- Secrets by call-home — delivered once over TLS via a single-use bootstrap token that burns on first use; sealed at rest, never in the provider console, never in logs
All of it provisions itself: pay → server created → installs itself → DNS + TLS → live, typically in about 5 minutes. From then on a reconciler converges every box back to the state it should be in, around the clock, and every state change is written to an append-only audit log.
Pricing
Flat for CPU. Metered for GPU. No surprises either way.
CPU plans are one flat number a month on enterprise hardware in EU data centers (Germany & Finland). That's the whole bill.
Activated
$49/mo
2 vCPU · 4 GB RAM · EU data center
- Full Growth OS stack, pre-installed
- Your subdomain with auto-TLS
- Daily encrypted backups
- Full data export, anytime
Managed
$99/mo
4 vCPU · 8 GB RAM · EU data center
- Everything in Activated
- Headroom for real automation load
- Self-healing reconciler, 24/7
- Every state change audit-logged
White Glove
$299/mo
Dedicated 8 vCPU · 16 GB+ · EU data center
- Everything in Managed
- Dedicated CPU — no noisy neighbors
- Priority support, direct line
- Migration help when you need it
| Feature | Activated$49/mo | Most popularManaged$99/mo | White Glove$299/mo |
|---|---|---|---|
| vCPU | 2 | 4 | 8 |
| Dedicated vCPU | — shared | — shared | ✓ no noisy neighbors |
| RAM | 4 GB | 8 GB | 16 GB+ |
| NVMe SSD storage | ✓ | ✓ | ✓ |
| Traffic | ✓ included | ✓ included | ✓ included |
| Auto-TLS on your subdomain | ✓ | ✓ | ✓ |
| Daily encrypted backups | ✓ | ✓ | ✓ |
| n8n included | ✓ | ✓ | ✓ |
| Postgres 16 + Redis 7 | ✓ | ✓ | ✓ |
| Self-healing reconciler | ✓ | ✓ | ✓ |
| Full export, anytime | ✓ | ✓ | ✓ |
| Priority support, direct line | Not included | Not included | ✓ |
| Choose a plan |
Every tier runs the identical stack. You're choosing headroom, not features held hostage. Daily backups ship offsite once storage is configured for your box.
AI Starter
$99/mo
+ $1.20 / GPU-hr
RTX 4090-class · dedicated GPU cloud (US)
- Inference, agents, ComfyUI
- Auto-suspend when idle
- Data persists on a network volume
- Metered to the GPU-hr you actually run
AI Pro
$199/mo
+ $2.99 / GPU-hr
A100 80GB · dedicated GPU cloud (US)
- Serious model serving and fine-tunes
- Auto-suspend when idle
- Data persists on a network volume
- Volume survives suspend and rescale
AI Scale
Custom
H100 · reserved 24/7
For teams that never want the meter to matter
- H100-class or reserved capacity
- Fixed monthly, sized to your workload
- Same stack, on dedicated capacity
- Direct line to the operator
Who runs on this
Built for people who outgrow shared platforms
Agencies running client automations
One box per client. Their workflows, data and credentials stay isolated on their own server — no shared-platform execution caps, no per-task pricing. If a client leaves, hand them their export archive. Clean.
Outbound and growth teams
Enrichment, scoring, sequencing, webhooks — workflows that hammer APIs all day. On your own n8n there's no metered execution bill creeping up behind you; the box just runs.
SaaS founders who want owned infra
Real Postgres, real Redis, standard open-source containers underneath, SSH on request. Nothing proprietary sits between you and your data — your archives restore on any standard container host.
AI builders on the GPU tiers
4090-class or A100 80GB pods, metered to the GPU-hour you actually run. Idle boxes auto-suspend — the meter stops while your data persists on its network volume.
Managed for you
Managed by machinery, not good intentions
This is what "we manage it" actually means — not a support inbox. These run on and around every box, on every plan.
Firewall by default
ufw ships default-deny with only 22, 80 and 443 open. Nothing else on your box is reachable from the internet.
Hardened SSH
Key-only auth, password logins disabled permanently, fail2ban banning brute-force attempts in the background.
Standard containers underneath
The whole stack is standard, portable open-source containers on named volumes. Inspectable, portable, nothing proprietary holding your data.
Health probes
The control plane probes your box over HTTPS around the clock. The reconciler converges every box back to the state it should be in and repairs DNS drift automatically.
Append-only audit log
Every state change your box goes through is written to an audit log that can't be edited after the fact.
Suspend-safe billing
Payment fails? Grace period, then the box suspends — data intact on its volumes. Payment recovers, box resumes. Destruction only ever happens through the gated cancel path.
How it works
From payment to live, with no human in the loop
-
You pay
Stripe confirms the payment. A signature-verified webhook queues your build — the only thing a person could do here is slow it down.
-
Your server is created
A machine spins up in an enterprise EU data center (or a pod on our dedicated GPU cloud) with your name on the label. Every step is idempotent — retries can never create a second server or a second charge.
-
It installs itself
Our zero-touch installer hardens the box and brings up the Growth OS stack. Then the box calls home over TLS to collect its secrets — credentials are sealed at rest and never ride in plain text.
-
DNS and TLS
Your subdomain points at your box and a real TLS certificate is issued automatically. No setup screen, no "propagation may take 48 hours."
-
Live — and it stays that way
Typically about 5 minutes after payment. From then on, a reconciler compares desired state against actual state around the clock and converges your box back to what it should be. Daily backups ship offsite without you asking.
The guarantee
Built, Not Rented.
Your box. Your data. Export anytime, leave anytime.
Cancel a CPU box and it is snapshotted in full before anything is destroyed — app, database, automations, files. Want your data in your hands, while you stay or on your way out? Email us and your archives are yours same-day, in standard formats — encrypted incremental snapshots, Postgres dumps, plain data volumes — that restore on any standard container host. No lock-in, no exit fee.
We don't do refund theater, because we don't need to: you always own your data, so you never have to ask for it back.
Start my boxFAQ
Straight answers
What is Growth OS?
What happens after I pay?
What's a GPU-hour, and what does auto-suspend mean?
Can I upgrade tiers?
What if I cancel?
Who manages updates and security?
Can I SSH in?
What regions do you run in?
Is my box shared with anyone?
How do backups work, exactly?
What happens if my payment fails?
Do you have an API? Can I self-manage?
Do I need to install or configure anything myself?
Pay at 11:17. Live by 11:22.
Pick a plan. The box builds itself. You own it from the first minute — and you can export it and leave whenever you want.