Small sites, VPNs and bots.
- CPU1 vCPU
- Memory1 GB
- Storage20 GB NVMe
- Port1 Gb/s
- TrafficUnmetered
- IPIPv4 + IPv6
VPS Hosting / n8n Hosting
Self-hosted n8n on KVMn8n hosting at RS Computers is a KVM VPS in Amsterdam, Dublin or Prishtina on which you run the free, self-hosted n8n community edition, so no one bills you per workflow execution. Every plan has NVMe storage, a dedicated IPv4 address plus IPv6, unmetered traffic and free weekly backups. n8n is not preinstalled: our step-by-step n8n guide sets it up with Docker, or we install it for you, quoted on request by Telegram or email.
Small sites, VPNs and bots.
Websites, apps and dev servers.
Busier apps, databases and Windows.
Production workloads on a 10 Gb/s port.
High-traffic sites and heavier databases.
Large apps, clusters and storage-heavy work.
Small sites, VPNs and bots.
Websites, apps and dev servers.
Busier apps, databases and Windows.
Production workloads on a 10 Gb/s port.
High-traffic sites and heavier databases.
Large apps, clusters and storage-heavy work.
Small sites, VPNs and bots.
Websites, apps and dev servers.
Busier apps, databases and Windows.
Production workloads on a 10 Gb/s port.
High-traffic sites and heavier databases.
Large apps, clusters and storage-heavy work.
Prishtina is sold out at the moment. The same plans and prices are available in Amsterdam and Dublin. Follow us on Telegram to hear when Kosovo capacity is back.
Every plan: NVMe storage, a dedicated IPv4 address plus IPv6, free weekly backups and full root access. Windows Server is available on VPS Mini and all VDS plans.
n8n Cloud plans count monthly executions. Self-hosted n8n on your own VPS is limited only by the server's memory and CPU.
Every plan has a dedicated IPv4 address plus IPv6, so webhook URLs stay stable and partners can allow-list your server.
Amsterdam and Dublin keep workflows and credentials inside the EU; Prishtina keeps them in Kosovo. Same plans and prices in all three.
Free weekly backups you restore from the client area, plus reboot, reinstall and a console in the same panel.
RS Computers offers n8n hosting as plain KVM servers that you control: a VPS or VDS in Amsterdam, Dublin or Prishtina, where you run the self-hosted n8n community edition. The six plans above cost the same in all three cities, with full root access, a panel to reboot, reinstall and open a console, and a 99% monthly uptime SLA.
n8n is not preinstalled. You install it with our n8n setup guide (Docker, Postgres and Caddy), or we set it up for you: installation and help are quoted on request, based on what you need, by Telegram or email. In return nobody counts your executions, and the server, the database and the encryption key for your credentials are yours.
n8n Cloud is run by the company behind n8n. The self-hosted version is the same core product, and n8n's hosting documentation says that without a licence key it runs as the free community edition. Where you run it is a separate decision.
| Option | Who runs it | Execution limits | Where the data lives | Fits best when |
|---|---|---|---|---|
| n8n Cloud (Starter, Pro) | n8n, including updates | 2,500 or 10,000 workflow executions a month, 5 or up to 50 at once | EU, on servers in Frankfurt | You want no server work at all |
| Self-hosted at home or in the office | You, on a PC or office server | None set by n8n; the hardware is the limit | Your own building | Learning and testing; webhooks need a public address and a machine that never sleeps |
| Self-hosted on a VPS | You, on a server in a data centre | None set by n8n; memory and CPU are the limit | The city you pick: Amsterdam, Dublin or Prishtina | Production workflows, public webhooks, a fixed IP for allow-lists |
The Cloud figures come from n8n's pricing page in October 2026, which counts one run of a whole workflow as one execution, however many steps it has, and warns that overage charges may apply past the quota. A workflow that polls an API once a minute runs 43,200 times in a 30-day month, more than the Pro allowance. On a VPS it only uses CPU time.
The price of self-hosting is work: updates, backups and locked-down logins are your job. The community edition also leaves out team features such as single sign-on, projects, environments, external secrets, Git version control and sharing workflows or credentials beyond the instance owner (n8n, Compare editions). Those need n8n's paid Business or Enterprise licence.
Self-hosted n8n is published under the Sustainable Use License. Its licence text allows use "only for your own internal business purposes or for non-commercial or personal use". n8n's licence FAQ reads that broadly: you can run your company on it, build automations for clients on your own instance as long as they cannot create or edit them, and charge for that work. It rules out hosting n8n as a service where clients build their own workflows, or white-labelling it.
n8n's documentation gives one firm number. The official Docker Compose install asks for at least 4 GB of RAM and 2 vCPUs, because its sandbox for AI-generated code runs Docker inside Docker. A plain n8n container without that sandbox is lighter. Past that, memory depends on your workflows: n8n's page on fixing memory issues names the amount of JSON data, binary file sizes, node count, Code nodes, manual test runs and executions running at once. Batching big jobs through sub-workflows keeps the peak down.
CPU is rarely the first wall. In n8n's benchmarks, a single instance with 4 GB of RAM and Postgres handled up to 220 executions per second of a simple webhook workflow. Concurrency control is off by default, so a burst of webhooks runs all at once and memory runs out first. vCPUs on our plans are shared, so leave headroom.
n8n uses SQLite unless you configure something else, and its docs advise Postgres once more than a handful of users or round-the-clock workflows are involved; supported versions are PostgreSQL 17 and 18, plus 16. By default n8n prunes finished executions older than 14 days or beyond the newest 10,000 (Manage execution data), so that history needs disk too.
Queue mode comes after one instance: a main process takes the editor and webhooks, Redis holds the queue, and workers run executions, 10 at a time each by default. n8n's queue mode docs advise against SQLite there, and workflows that keep binary files need S3 external storage, which the community edition does not include.
| Workload | What drives the size | Where to start |
|---|---|---|
| Learning n8n, a few personal workflows | A plain n8n container on SQLite | VPS Nano or VPS Micro |
| A small business: website forms, shop orders, CRM sync | n8n, Postgres and a reverse proxy side by side, plus webhook bursts | VPS Micro, or VPS Mini to match n8n's Compose requirement |
| The full Compose stack with the AI sandbox, AI agents calling hosted models, several editors | n8n's documented minimum of 4 GB and 2 vCPUs | VPS Mini or VDS Small |
| Large files, big data syncs, queue mode with Redis and workers | Memory per concurrent execution and extra processes | VDS Small or VDS Medium |
| Small local models with Ollama next to n8n | The model has to fit in memory, beside everything else | VDS Medium or VDS Large |
The plan cards above list each plan's exact vCPU, memory and disk.
An automation server sees customer records, invoices and the API keys for every service it touches. n8n Cloud stores data in Frankfurt, inside the EU, according to its pricing FAQ, so the Cloud is not a GDPR problem in itself. Self-hosting changes who holds the data: on an RS Computers VPS, workflows, credentials and execution history sit on a disk in the city you chose, and a hosting provider is normally your processor under Article 28 of the GDPR.
Two details often missed. Self-hosted n8n sends anonymous telemetry, update notices and template requests to n8n's servers by default, and the Isolate n8n page lists the settings that switch each one off. And data leaves your server whenever a workflow sends it on: an AI node calling OpenAI hands that text to another company, which belongs in your records too.
Support answers on Telegram and at support@rscomputers-ks.com, Monday to Friday from 09:00 to 17:00, and the NOC watches the platform 24/7. There is no phone line. To compare everything in one place, see all VPS and VDS plans.
Yes. n8n's hosting docs say every self-hosted install without a licence key runs as the free community edition, with no execution cap set by n8n. The Sustainable Use License allows internal business, personal and non-commercial use, but not hosting n8n for clients to build their own workflows. What you pay for is the server: on RS Computers, a flat plan in Amsterdam, Dublin or Prishtina.
It depends on the setup. n8n's official Docker Compose install asks for at least 4 GB of RAM and 2 vCPUs, because it includes a sandbox for AI-generated code. A plain n8n container on SQLite with a few light workflows needs less. Large JSON payloads, files, Code nodes and many executions at once raise memory use, so size the server for your busiest workflow.
n8n Cloud suits teams that want no server work: n8n handles updates, and data stays in Frankfurt. Its Starter and Pro plans include 2,500 and 10,000 monthly executions, per n8n's pricing page in October 2026. Self-hosting on a VPS removes the execution count and gives you a fixed IP and your choice of country, but you handle updates, security and backups yourself.
Not strictly, but it is the sensible route. n8n's docs present Docker Compose as the production install and say npm installs are deprecated from n8n 3.0. Docker keeps n8n, Postgres and the reverse proxy in separate containers, and updates become a pull and restart. Our n8n setup guide walks through the whole Docker install on an RS Computers VPS.
n8n's update page advises updating at least once a month, so you never jump many versions at once. On 5 October 2026 the newest stable release on GitHub was 2.42.3. With Docker you pull the new image and recreate the container; our setup guide shows the commands. Read the release notes for breaking changes and back up the database first.
Yes. AI Agent nodes that call OpenAI, Anthropic or another hosted model are light work for the server, since the model runs elsewhere. Running a model locally with Ollama is different: the model must fit in memory next to n8n, and small models are what a VPS handles well. n8n's own AI starter kit is meant for proofs of concept and needs hardening before production.
The hosting location helps, but compliance depends on how you process the data. A server in Amsterdam or Dublin keeps n8n's database inside the EU; Prishtina keeps it in Kosovo, outside the EU. Self-hosted n8n sends anonymous telemetry to n8n by default, which you can switch off. Every external service your workflows call also receives data, so list those providers in your records.
No. RS Computers servers come with a clean Linux image such as Ubuntu or Debian, and n8n is not preinstalled. You install it with our step-by-step guide, which uses Docker, Postgres and Caddy, or ask us to set it up for you, quoted on request. Message us on Telegram, Monday to Friday from 09:00 to 17:00. The server is deployed automatically after payment, usually within minutes.
Ask on Telegram and a person from our team answers: Mon to Fri 09:00 to 17:00, with a 24/7 NOC for emergencies.