Hosting for AI services

n8n, Open WebUI, Flowise and Ollama: your data stays on your server

Pick a configuration Order in the bot

Price: from $2.99/mo
Locations: 4 in Europe
Apps: 5 apps in one click
Rental: from 1 day

Configurations for the job

Six steps from small to large, with resources and price matched to the job.

NLDEPLEE

01
Automations on n8n
Workflows, webhooks and integrations
Why these resources
CPU
1 vCPU
Memory
2 GB DDR5
Disk
20 GB NVMe
$3.29/mo
$0.12 for 24 hours
Order in the bot
$4.48/mo
$0.16 for 24 hours
Order in the bot
$2.99/mo
$0.11 for 24 hours
Order in the bot
$3.89/mo
$0.14 for 24 hours
Order in the bot
02
Chat with models via API
Open WebUI or Flowise on top of external models
Why these resources
CPU
2 vCPU
Memory
4 GB DDR5
Disk
30 GB NVMe
$6.58/mo
$0.24 for 24 hours
Order in the bot
$8.97/mo
$0.32 for 24 hours
Order in the bot
$5.98/mo
$0.21 for 24 hours
Order in the bot
$7.77/mo
$0.28 for 24 hours
Order in the bot
03Recommended
Local 7–8B model
Ollama on the CPU: 7–8B models with 4-bit quantization
Why these resources
CPU
4 vCPU
Memory
8 GB DDR5
Disk
40 GB NVMe
$11.38/mo
$0.41 for 24 hours
Order in the bot
$15.52/mo
$0.55 for 24 hours
Order in the bot
$10.35/mo
$0.37 for 24 hours
Order in the bot
$13.45/mo
$0.48 for 24 hours
Order in the bot
04
Knowledge base with a local model
RAG: a 7–8B model, a vector database and n8n
Why these resources
CPU
6 vCPU
Memory
12 GB DDR5
Disk
80 GB NVMe
$20.90/mo
$0.75 for 24 hours
Order in the bot
$28.50/mo
$1.02 for 24 hours
Order in the bot
$19.00/mo
$0.68 for 24 hours
Order in the bot
$24.70/mo
$0.88 for 24 hours
Order in the bot
05
13–14B model
Ollama with 13–14B models
Why these resources
CPU
8 vCPU
Memory
16 GB DDR5
Disk
80 GB NVMe
$22.76/mo
$0.81 for 24 hours
Order in the bot
$31.04/mo
$1.11 for 24 hours
Order in the bot
$20.69/mo
$0.74 for 24 hours
Order in the bot
$26.90/mo
$0.96 for 24 hours
Order in the bot
06
Models up to 32B
Ollama with models up to 32B, 4-bit quantized
Why these resources
CPU
12 vCPU
Memory
32 GB DDR5
Disk
150 GB NVMe
$44.11/mo
$1.58 for 24 hours
Order in the bot
$60.15/mo
$2.15 for 24 hours
Order in the bot
$40.10/mo
$1.43 for 24 hours
Order in the bot
$52.13/mo
$1.86 for 24 hours
Order in the bot

Need a different size? Build a configuration or pick a ready-made plan.

Similar servers elsewhere

Similar plans from other providers next to our steps in Poland.

Our step 03

Local 7–8B model

4 vCPU · 8 GB · 40 GB NVMe

$10.35/mo

Aéza
WAWs-3 · Warsaw
4 vCPU · 8 GB · 120 GB NVMe
They offer more: disk
$26.70/mo
Our step is cheaper by $16.35 a month, or 61%
Source: aeza.netchecked October 17, 2026on their site €23.73/mo, ECB rate on October 16, 2026: $1.1253 per €1
VDSina
Стандартный, 4 core · 8 Gb
4 vCPU · 8 GB · 150 GB
They offer more: disk
$28.70/mo
Our step is cheaper by $18.35 a month, or 64%
Source: vdsina.ruchecked October 17, 2026on their site RUB 2,400/mo, Central Bank of Russia rate on October 17, 2026: 83.6177 ₽ per $1
Hetzner
CPX32 · Germany, Finland
4 vCPU · 8 GB · 160 GB NVMe
They offer more: disk
$41.99/mo
excl. IPv4 and VAT
Our step is cheaper by $31.64 a month, or 75%
Source: docs.hetzner.comchecked October 17, 2026
DigitalOcean
Basic Droplet Regular, 4 vCPU · 8 GiB
4 vCPU · 8 GB · 160 GB SSD
They offer more: disk
$48.00/mo
Our step is cheaper by $37.65 a month, or 78%
Source: digitalocean.comchecked October 17, 2026

Prices are taken from the providers' pricing pages on the check date, per month, without prepayment discounts; rubles and euros are converted to US dollars at the official rate (Central Bank of Russia and ECB respectively). Our price is the step from the list above, with the server in Poland. Resources don't match one to one: where another provider gives more of a resource, the row says so. Providers change their prices; the current price is on their site, linked in each row.

Why these resources

Minimums come from the apps' official requirements; our estimates are marked.

01Automations on n8n1 vCPU · 2 GB · 20 GB · estimate
n8n has no official minimum. The library lists 1 vCPU and 2 GB for n8n, plus room for SQLite and execution logs.
- n8n: hosting
02Chat with models via API2 vCPU · 4 GB · 30 GB · estimate
The Open WebUI and Flowise images weigh several gigabytes, and local embeddings take memory too.
03Local 7–8B model4 vCPU · 8 GB · 40 GB · estimate
A 7–8B model with 4-bit quantization takes about 5 GB on disk and in memory, plus context and the OS.
04Knowledge base with a local model6 vCPU · 12 GB · 80 GB · estimate
A 7–8B model, a vector database and n8n on one server. Our estimate.
0513–14B model8 vCPU · 16 GB · 80 GB · estimate
A 13–14B model with 4-bit quantization takes about 8–9 GB; the rest is for context and the OS.
06Models up to 32B12 vCPU · 32 GB · 150 GB · estimate
A model of up to 32B in 4 bits is about 20 GB, plus context. 150 GB leaves room for several models.

What's included

n8n, Ollama, Open WebUI and Flowise are installed with the order, at the “Add-ons” step
Rich Agent in the bot runs commands on the server and asks for confirmation before every action
No sign-up, no ID checks
Rent from 24 hours
Up and running in ~60 seconds
Dedicated IPv4 with a reputation check
Intrusion protection
Up to 25 Gbps port
Migration across 4 locations
Public API
Shared access

One-click apps

Tick the ones you need when ordering, and they are installed together with the server.

Full library

Install over SSH

root@vps:~# command -v docker >/dev/null || curl -fsSL https://get.docker.com | sh; docker volume create n8n_data && docker run -d --name n8n --restart unless-stopped -p 5678:5678 -e GENERIC_TIMEZONE=Europe/Moscow -e TZ=Europe/Moscow -v n8n_data:/home/node/.n8n docker.n8n.io/n8nio/n8n

SystemUbuntu 24.04 · x86_64ok
Sourcegithub.com/n8n-io/n8nok
Version2.41.4ok
Port5678/tcplisten

✓ n8n 2.41.4 installed · about 2 minutes

root@vps:~#

Install command from the library. Output is shortened and may differ on your server.

n8n
Workflow automation builder
AI / AutomationFree
Ollama
Run LLMs locally
AI / AutomationFree
Open WebUI
Web UI for LLMs
AI / AutomationFree
Flowise
AI agent builder
AI / AutomationFree
Python
Language for bots and scripts
BotsFree

Where to host

The price of the “Local 7–8B model” step in each data center.

Locations and the price of the “Local 7–8B model” step
Location	CPU	Port	Per month
NetherlandsAmsterdam	Ryzen 9950x / Intel 14900K	10 Gbps	$11.38
GermanyFrankfurt	Ryzen 9950x / Ryzen 9950x3d	25 Gbps	$15.52
PolandWarsaw	Ryzen 9950x / Ryzen 7950x	5 Gbps	$10.35
EstoniaTallinn	Ryzen 9950x / Ryzen 9950x3d	10 Gbps	$13.45

Check the latency to each data center before you order: Looking Glass · Speedtest

Pay only for what you need

Filled is what the job uses, empty is overpayment. Prices in Poland, per month.

01Automations on n8n
Needs: 1 vCPU · 2 GB · 20 GB
Ready-made plan PL-R9-2: 1 vCPU · 3 GB · 30 GB
CPU
1 of 1 vCPU
Memory
2 of 3 GB
Disk
20 of 30 GB
$2.99/mo
instead of $3.19 for the ready-made plan
6% cheaper
02Chat with models via API
Needs: 2 vCPU · 4 GB · 30 GB
Ready-made plan PL-R9-4: 4 vCPU · 4 GB · 100 GB
CPU
2 of 4 vCPU
Memory
4 of 4 GB
Disk
30 of 100 GB
$5.98/mo
instead of $10.20 for the ready-made plan
41% cheaper
03Local 7–8B model
Needs: 4 vCPU · 8 GB · 40 GB
Ready-made plan PL-R9-5: 4 vCPU · 8 GB · 150 GB
CPU
4 of 4 vCPU
Memory
8 of 8 GB
Disk
40 of 150 GB
$10.35/mo
instead of $13.30 for the ready-made plan
22% cheaper
0513–14B model
Needs: 8 vCPU · 16 GB · 80 GB
Ready-made plan PL-R9-8: 8 vCPU · 16 GB · 225 GB
CPU
8 of 8 vCPU
Memory
16 of 16 GB
Disk
80 of 225 GB
$20.69/mo
instead of $21.04 for the ready-made plan
2% cheaper

AI services FAQ

No answer here? Message @ExpressHost_SupportBot.

Is there a GPU?

No, models run on the Ryzen 9950x CPU. For 7–8B models with 4-bit quantization we plan 4 vCPU, 8 GB RAM, 40 GB NVMe, for 13–14B, 8 vCPU, 16 GB RAM, 80 GB NVMe. Generation speed depends on the model and the context length.

Which model can the server handle?

Memory is the guide: a 7–8B model in 4 bits takes about 5 GB, 13–14B about 9 GB, up to 32B about 20 GB, plus context and the OS. That's our estimate based on model file sizes.

Does my data go to external services?

With a local model in Ollama, requests are processed on your server. If you connect an external API in Open WebUI, Flowise or n8n, the request text goes to that provider.

What resources does n8n need?

n8n has no official minimum. For personal workflows on SQLite we plan 1 vCPU, 2 GB RAM, 20 GB NVMe ($2.99 a month); that's our estimate. When you run more executions, change the plan with the “Plan” button in the bot.

AI services on Ryzen and NVMe — in 4 steps.

Choose a plan Buy in Telegram

Hosting for AI services

Automations on n8n

Chat with models via API

Local 7–8B model

Knowledge base with a local model

13–14B model

Models up to 32B

n8n

Ollama

Open WebUI

Flowise

Python

AI services FAQ

Is there a GPU?

Which model can the server handle?

Does my data go to external services?

What resources does n8n need?