Your stack, your server
Run n8n, Nextcloud, Ollama and anything Docker ships on a Linux VPS you control, instead of a stack of SaaS subscriptions that meter your usage and hold your data. From $4.50 per month across 8 cloud providers, payable in crypto or PayPal.
One small server replaces a pile of subscriptions
Your data stays yours
Workflows, files, chat logs and model prompts live on a server you hold root on, not in a vendor account that can change terms, raise prices or close. Move providers whenever you like; it is your disk image.
Flat cost, no per-seat meter
SaaS pricing scales with users, executions and tokens. A VPS costs the same whether your n8n runs ten workflows or ten thousand. One mid-size box happily runs several apps side by side under Docker.
Runs where you want it
Pick from 126 locations across 8 cloud providers, all billed on one prepaid Initech balance. Keep your stack in the jurisdiction you prefer and next to the services it talks to.
What are you self-hosting?
n8n
Self-hosted n8n runs every workflow your server can handle with no per-execution billing, and your credentials for connected services stay on your box. A small VPS covers most automation loads; add RAM if you process large payloads. Our step-by-step guide takes you from blank server to running instance.
Read the n8n setup guide →Docker and Coolify
Anything published as a container image runs on a plain Linux VPS: databases, dashboards, game servers, internal tools. Prefer a deploy UI over the command line? Install Coolify or a similar self-hosted PaaS on the same box and push apps to it like a private Heroku.
Read the Docker guide →Ollama and open LLMs
Ollama makes running open models a one-line install, and self-hosting means prompts never leave your server. Be realistic about sizing: a CPU VPS with RAM to fit the model handles small models at modest speed, which suits background and batch jobs. For interactive speed or larger models, use our NVIDIA GPU line.
GPU VPS with NVIDIA A16 and A40 →ComfyUI and image generation
Diffusion workloads want a GPU; CPU-only generation is too slow to be useful. Rent a GPU VPS by the month, install ComfyUI or Automatic1111 under your own login, and keep your models, LoRAs and outputs on your own disk instead of a shared platform.
See GPU plans →Nextcloud
Your own file sync, calendar and document editing on hardware you choose, in the country you choose. Storage is the sizing question: pick a plan with the disk you need, or a configurable line where you set the disk size directly. Our guide covers install through first sync.
Read the Nextcloud guide →AI agents and bots
Agent frameworks and API-driven bots need somewhere permanent to live: a stable IP for webhooks, cron for schedules, and a process that survives your laptop closing. A small Linux VPS is that home. Pair it with self-hosted n8n for triggers, or point it at whatever model API you already use.
Linux VPS from $4.50/mo →All servers are unmanaged: you install and run your own software under your own licenses, with our guides to walk you through the common stacks. Standard acceptable use policy applies.
Pick a line, resize when you outgrow it
| Line | From | Good for | Locations | |
|---|---|---|---|---|
| Initech Linux (configurable) | $4.50/mo | Pick exact vCPU, RAM and disk | 1 | 部署 → |
| Hetzner Cloud | $9/mo | Most RAM per dollar in the EU | 5 | 部署 → |
| Vultr Cloud VPS | $10/mo | Widest choice of cities | 32 | 部署 → |
| NVIDIA GPU line (A16, A40) | Ollama at speed, ComfyUI, CUDA | See GPU plans → |
Retail USD pricing at the smallest configuration and lowest-priced region; the order form shows exact figures. Five more Linux lines (Amazon EC2, Lightsail, DigitalOcean, Google Cloud, Alibaba Cloud) are on the full pricing table, 126 locations combined, all on one prepaid balance. See all lines
Blank server to running stack
Unmanaged does not mean unsupported reading. Our knowledgebase walks through the common self-host stacks step by step:
Self-hosting questions
What size VPS do I need for n8n?
n8n itself is light, and an entry plan runs typical personal and small-team automation comfortably. What grows the requirement is workflow volume and payload size, so start small and resize when you see memory pressure. Self-hosting removes per-execution billing entirely.
Can I run Ollama or other local LLMs on a CPU VPS?
Yes, with honest expectations. The model must fit in RAM, and CPU inference of small models is fine for background jobs, summarization queues and experiments but not for snappy chat. For interactive use or larger models, rent a server from our NVIDIA GPU line instead.
Do you install or manage the applications?
No. These are unmanaged servers with full root access: you install your stack, you administer it, and our step-by-step guides cover the common setups. If you want hands-off SaaS, self-hosting is the wrong tool; if you want control, this is it.
Is self-hosting actually cheaper than SaaS?
Often, but not automatically. One $10 server running n8n, Nextcloud and a couple of containers replaces several per-user subscriptions, and heavy automation users save the most because executions are free. Factor in your own admin time; the guides shrink it but do not remove it.
Can I pay with crypto?
Yes. Top up your prepaid balance with Bitcoin, Lightning, Monero and other coins, or use PayPal. Registration takes an email address and a country; no card is required, and one balance covers servers on all 8 providers.
Move one app off SaaS this week
Linux VPS from $4.50/mo, GPU servers for the AI pieces, guides for the setup. Online in minutes.
Deploy your serverGPU VPS · Pay with Bitcoin · Compare all providers · Setup guides
