Most people who want an AI agent running day and night need a small rented server, not a GPU. The agent sends its requests to an AI model over the internet. The server only has to stay on and hold the agent's own software. If you would rather not look after a server at all, a managed plan does the upkeep for you, at a higher monthly price.
This guide covers the three best-known agent apps: n8n, OpenClaw and Hermes Agent. It compares the routes from the apps' own documentation and each host's pages, read on 27 October 2026.
Who this is for
- you want n8n, OpenClaw or Hermes Agent running all the time, not only while your laptop is on
- you are choosing between a managed plan, a server with the app pre-installed and a plain server
- you want to know what it really costs each month, model included
Not for you if you want to run the AI model itself on your own hardware: see where to run an open-source LLM.
Key numbers
- 640 MiB the memory n8n Cloud runs its Starter plan on, by n8n's documentation Official page,
- 2 GB the top of the memory range in n8n's own illustrative sizing table, which starts at 320 MB Official page,
- 1 GB of memory with swap in OpenClaw's DigitalOcean guide, or 2 GB if it runs out Official page,
- 18789 port OpenClaw's Gateway port, which its guide reaches through an SSH tunnel, never the open internet Official page,
Which route fits you
There are four ways to keep an agent running, and they differ mainly in how much upkeep you take on. The questions below point to the right one for you:
Where should your agent run?
Question 1 Will you update and look after a server yourself?
- No: a managed plan
- Yes, or I am willing to learn: go to question 2
Question 2 Do you want the app installed for you when you order?
- Yes: a one-click server
- No, I want full control: go to question 3
Question 3 Must it run when your own computer is off?
- You and your apps (Telegram, e-mail, webhooks): connects to Your agent on a server (messages and webhooks).
- Your agent on a server (n8n, OpenClaw or Hermes Agent): connects to AI model (prompts over HTTPS).
- AI model (through an API, paid per token)
Managed agent plans: the host keeps it running
On a managed plan the host installs the agent, applies updates and keeps backups, so you only configure the agent itself. You still bring your own key for the AI model in most cases and pay that provider separately.
- Cloudways runs OpenClaw or Hermes Agent for you, with managed SSL, automated daily backups and updates it tests before release. Its public pages name no regions, so its prices are not in our records yet
- Hosting.com sells a fully managed server that starts with OpenClaw installed; it patches, monitors and backs up the server
- Hostinger's AI automation apps plan covers OpenClaw, Hermes Agent, n8n or Paperclip, and you can switch between them
- for n8n alone, n8n Cloud is n8n run by its own makers in Frankfurt, priced by workflow executions
- ScalaHosting sets up n8n for you on a managed cloud server
| Plan | Price | New-customer offer | Link |
|---|---|---|---|
| Hosting.com Managed Server 4GB (OpenClaw pre-installed) 2 vCPU, 4 GB | $36/month 1-year term, paid up front; introductory first-year price, then $40 checked 25 Oct 2026 | No offer | Visit |
| n8n Cloud Starter | €20/month yearly plan, paid up front for 12 months and not refundable; works out at this price per month checked 25 Oct 2026 | Free trial of n8n Cloud with Pro-level features and no card needed checked 29 Oct 2026 | Visit |
| n8n Cloud Pro | €50/month yearly plan, paid up front for 12 months and not refundable; works out at this price per month (the 10,000-executions-a-month option) checked 26 Oct 2026 | Free trial of n8n Cloud with Pro-level features and no card needed checked 29 Oct 2026 | Visit |
n8n Cloud prices are in euros and Hosting.com's in US dollars. Model usage is extra on Hosting.com.
One-click agent servers: installed for you, looked after by you
Several hosts install an agent on a new virtual private server (VPS) when you order it. You get a working agent in minutes, but the server is yours to update and secure from then on.
- Hostinger's order page can install n8n, OpenClaw, Hermes Agent or Ollama
- DigitalOcean offers one-click images for OpenClaw, Hermes Agent and n8n
- Kamatera's marketplace has images for OpenClaw, n8n and Hermes Agent
- Contabo's order page sets up n8n, OpenClaw or Hermes Agent at no extra cost
OpenClaw's own DigitalOcean guide suggests a clean Ubuntu image instead of third-party marketplace images you have not reviewed. A one-click image saves time, but you cannot always see what it installed, so check it.
For OpenClaw and Hermes Agent, our guides to OpenClaw hosting and Hermes Agent hosting compare what each host's image sets up.
For n8n, our n8n hosting guide compares what each host's ready-made server sets up. Our five picks of servers for n8n match a plan to each need.
| Plan | Price | New-customer offer | Link |
|---|---|---|---|
| Hostinger KVM 1 1 vCPU, 4 GB | $6.49/month 24-month term, paid up front; Black Friday first-term price, then $11.99 checked 25 Oct 2026 | Black Friday prices on VPS plans and AI automation apps, for the first subscription term only checked 30 Oct 2026 | Visit |
| Hostinger KVM 2 2 vCPU, 8 GB | $8.79/month 24-month term, paid up front; Black Friday first-term price, then $14.99 checked 25 Oct 2026 | Black Friday prices on VPS plans and AI automation apps, for the first subscription term only checked 30 Oct 2026 | Visit |
| DigitalOcean Basic Droplet, 2 vCPUs, 4 GiB 2 vCPU, 4 GB | $24/month billed per second (60-second minimum), capped at this monthly price · VAT not stated checked 25 Oct 2026 | New accounts get a $5 credit that lasts 90 days and covers every product except SaaS add-ons checked 29 Oct 2026 | Visit |
| Kamatera Cloud server Type B (General Purpose), 2 cores, 4 GB 2 vCPU, 4 GB | $39/month monthly, no long-term contract · VAT not stated checked 25 Oct 2026 | A 30-day trial of one cloud server in a configuration worth up to $100 checked 29 Oct 2026 | Visit |
| Contabo Cloud VPS 4 4 vCPU, 8 GB | €4.40/month 24-month term, paid up front; the price applies to the first 24 months and to the EU region in Germany; other regions cost more each month checked 25 Oct 2026 | No offer | Visit |
Some of the lowest prices need a long term paid in advance; the term is under each price.
A plain VPS with Docker: the most control
On a plain server you install the agent yourself, usually with Docker, and you decide every setting. It gives you the most control and teaches you the most. Every update and every security setting is your job, though.
| Plan | Price | New-customer offer | Link |
|---|---|---|---|
| Hosting.com Self Managed Linux 4GB 2 vCPU, 4 GB | $6.49/month 3-year term, paid up front; introductory price for the first three years, then $19 checked 25 Oct 2026 | No offer | Visit |
| Liquid Web Cloud VPS (General), 2 vCPUs, 4 GB 2 vCPU, 4 GB | $17/month monthly · VAT not stated checked 25 Oct 2026 | No offer | Visit |
| UpCloud Starter, 2 CPU cores, 4 GB 2 vCPU, 4 GB | €12/month billed by the starting hour, at most 28 days a month; no minimum term checked 25 Oct 2026 | No offer | Visit |
| Cherry Servers Cloud VPS 4 (Gen 2) 4 vCPU, 4 GB | €9/month monthly, no long-term contract checked 25 Oct 2026 | No offer | Visit |
| Infomaniak VPS Lite 1 vCPU / 2 GB RAM (Linux) 1 vCPU, 2 GB | €2.70/month monthly rate of a 12-month commitment, paid in one payment for the year; with Linux (Windows costs more) checked 25 Oct 2026 | No offer | Visit |
Prices in euros and US dollars as each provider shows them; we do not convert them.
For n8n, our step-by-step guide to self-hosting n8n with Docker covers HTTPS, backups and safe updates.
Size the server from the app's own figures. For n8n, our guide to how much RAM and CPU n8n needs sets them out. OpenClaw's DigitalOcean guide starts at 1 GB of memory with swap.
At home on your own computer: when that is enough
OpenClaw installs on macOS, Linux and Windows, and Hermes Agent has a desktop app for macOS and Windows. Running an agent on a computer you already own costs nothing extra, but it stops whenever that computer sleeps, loses power or drops its connection.
Home is fine for trying an agent out. Once other people or other systems depend on it, a rented server is the safer choice. Our guide to OpenClaw on a Mac mini or a VPS weighs the two.
Do you need a GPU?
Not if the agent uses a model through an API, which is how most agents run. OpenClaw works with hosted models such as Claude and GPT, and Hermes Agent with your own model provider or Nous Portal. The server then does little more than wait for messages and pass them on.
You need a GPU only when the model itself runs on the same machine. Small models can run on an ordinary server's processor, as our guide to running Ollama without a GPU shows. Larger ones need a GPU server, covered in where to run an open-source LLM.
What it costs each month
An agent's monthly bill has three parts, and the server is often the smallest:
- the server or managed plan, a fixed price per month (in the boxes above)
- the AI model, paid per token and rising with use
- extras people forget: backups where they are not included, a domain name, and longer terms that renew at a higher price
Our guide to what OpenClaw or Hermes Agent costs each month has a calculator for the server and the model together. For n8n, our comparison of n8n Cloud and self-hosting does the same for three example workloads.
Keep it safe
An agent can read your messages, files and accounts, so a careless set-up is a real risk. Three habits cover most of it:
- keep the agent's control panel off the open internet; OpenClaw's guide reaches its Gateway through an SSH tunnel or Tailscale
- give the agent its own accounts and keys, with spending limits on the model API
- update the agent and the server's system regularly, and keep a backup you have restored at least once
Our guides to securing OpenClaw on a server and securing a self-hosted n8n server go step by step.
The routes side by side
A managed plan costs more because the host does the upkeep; every other route leaves it to you.
| Route | What is left for you | Who installs updates | Example monthly price |
|---|---|---|---|
| Managed plan | set up the agent itself | the host | $36 a month (checked 25 October 2026) |
| One-click server | update and secure the server | you | $8.79 a month (checked 25 October 2026) |
| Plain VPS with Docker | install, update and secure everything | you | €12 a month (checked 25 October 2026) |
| Your own computer | keep it switched on | you | no rent; electricity only |
How we checked
We read the apps' documentation and each host's pages on 27 October 2026. Plan facts and prices come from our weekly price checks.
Our own measurements will be added here, with their dates, when our test bench runs them.
Mistakes to avoid
Renting a GPU for an agent that calls a model API
Instead: Start with a small server; the model runs at the API provider.
Leaving the agent's control panel open to the internet
Instead: Keep it on the server and reach it through an SSH tunnel or Tailscale.
Comparing a 24-month first-term price with a monthly one
Instead: Read the term and the renewal price under each price.
Giving the agent your main accounts and an uncapped API key
Instead: Use separate accounts and set a spending limit.
Trusting a one-click image you have not checked
Instead: Check what it installed, or start from a clean image.
Other questions
Can I run an AI agent for free?
On a computer you already own, yes, apart from the model's usage costs. The n8n Cloud pricing page also lists a free trial without a card, capped at 1,000 executions.
How much memory does an AI agent need?
n8n's own sizing table gives 320 MB to 2 GB, and OpenClaw's DigitalOcean guide starts at 1 GB with swap. Files, many parallel runs or a local model need more.
Is a managed plan worth it?
If you would not otherwise update and back up the server, yes. That upkeep is what the higher price buys.
Sources and test notes
- n8n documentation: OEM deployment prerequisites (illustrative sizing), read
- n8n Cloud: memory on each plan, read
- OpenClaw documentation: install, read
- OpenClaw documentation: Linux server and VPS, read
- OpenClaw documentation: DigitalOcean, read
- OpenClaw home page, read
- Hermes Agent home page (Nous Research), read
- Cloudways managed AI agents, read
- Hosting.com OpenClaw hosting, read
- Hostinger VPS hosting, read
- Hostinger AI automation apps, read
- ScalaHosting self-hosted n8n cloud, read
- n8n pricing, read
- Kamatera pricing, read
- Contabo VPS, read
Every page read on 27 October 2026; plan facts and prices from our records.
Guides in this section
-
Ollama on a CPU serverHow fast small models run on a processor alone, from LocalScore's public results, what fits in 8, 16 and 24 GB, and how to set up Ollama safely.
-
n8n hostingn8n Cloud, managed n8n, a one-click server or Docker on your own VPS: what each one sets up, who installs updates and what it costs per month.
-
OpenClaw hostingManaged OpenClaw, one-click servers or your own VPS: what each one sets up, how much memory OpenClaw's guide and the hosts suggest, and the monthly cost.
-
Hermes Agent hostingWhat Hermes Agent needs from a server, by its own documentation, which hosts install or run it for you, and what each one-click image sets up.
-
OpenClaw and Hermes Agent costsA calculator that adds this week's server prices to a model API at your own usage, with what drives the token bill and when a subscription can pay.
-
Hermes Agent or OpenClawHermes Agent or OpenClaw: chat apps, security model, update pace and server needs, compared from both projects' own documentation and GitHub data.
-
Agent platformsDify, Langflow and n8n on your own server: what each is for, licences, the server each needs, install steps, and what to do now that Flowise has ended.
-
Remote MCP serversHost a remote MCP server on a small VPS or with n8n: the transport, step-by-step set-up with the official SDK, authentication and a security checklist.