n8n · Workflow automation · AI agents

VPS for n8n and AI Agents

Self-host n8n on a Virteche Cloud VPS in Ashburn, Virginia or Dallas, Texas: a KVM server with full root access, NVMe storage and its own IPv4 address for webhooks, built a few minutes after payment. Plans start at $37.50 a month with 6 GB of memory, several times what a typical n8n setup uses.

From Order to First Workflow

  1. 1

    Order a VPS

    Pick a plan and location below. Ubuntu or Debian is a good choice; the server is ready a few minutes after payment.
  2. 2

    Install Docker and n8n

    Follow the n8n self-hosting guide: Docker Compose, a data volume, and HTTPS in front with your own domain.
  3. 3

    Lock it down

    Allow only the web ports in the free host firewall, and keep n8n's editor behind its login. Blocked traffic never reaches the server.
  4. 4

    Connect your AI provider

    Add your OpenAI, Anthropic or other API key as an n8n credential and build agents with n8n's AI nodes.

VPS Plans and Prices

Every plan runs n8n comfortably. Pick the location closest to the services your workflows call.

VPS plans for n8n with monthly prices
PlanLocationCPU · RAM · DiskTransfer / PortPriceOrder
BronzeAshburn, VA3 vCPU · 6 GB · 75 GB NVMe2 TB / 10 Gbps$37.50/moConfigure
BronzeDallas, TX3 vCPU · 6 GB · 75 GB NVMe2 TB / 50 Gbps$37.50/moConfigure
SilverAshburn, VA4 vCPU · 8 GB · 100 GB NVMe4 TB / 10 Gbps$46.50/moConfigure
SilverDallas, TX4 vCPU · 8 GB · 100 GB NVMe4 TB / 50 Gbps$46.50/moConfigure
GoldAshburn, VA6 vCPU · 12 GB · 150 GB NVMe6 TB / 10 Gbps$60.00/moConfigure
GoldDallas, TX6 vCPU · 12 GB · 150 GB NVMe6 TB / 50 Gbps$60.00/moConfigure
DiamondAshburn, VA8 vCPU · 16 GB · 200 GB NVMe8 TB / 10 Gbps$73.50/moConfigure
DiamondDallas, TX8 vCPU · 16 GB · 200 GB NVMe8 TB / 50 Gbps$73.50/moConfigure
PlatinumAshburn, VA12 vCPU · 24 GB · 300 GB NVMe10 TB / 10 Gbps$100.00/moConfigure
PlatinumDallas, TX12 vCPU · 24 GB · 300 GB NVMe10 TB / 50 Gbps$100.00/moConfigure

What You Get

memory

KVM with root access

A full virtual machine with its own kernel: Docker, Compose and anything n8n's community nodes need run as on hardware.

webhook

A fixed public IPv4

Webhooks reach your server directly, and services that allowlist IPs can trust one address.

shield

Host firewall, free

Rules run on the host, so ports you block stay blocked whatever Docker does inside the server.

hard_drive

NVMe storage

n8n's database and execution history on local NVMe drives.

schedule

Always on

Schedules and webhooks keep running when your own computer is off.

trending_up

Resize later

Move to a larger plan from the dashboard as your workflows grow.

Good to Know Before You Order

  • infoThere is no GPU. AI models run at your provider; for running a model yourself, see high-memory dedicated servers.
  • infoOutgoing email on port 25 is blocked. Have n8n send mail through your email provider on port 587 or 465.
  • infoIPv4 only for now; IPv6 is not offered yet.
  • infoBackups are an add-on. Without it, copy the n8n_data volume to a storage box on a schedule.

Questions

How big a server does n8n need?expand_more

n8n itself is light: 1 to 2 GB of memory is enough for most personal and small-team setups. Workflows that handle large files or many runs at once need more, so watch docker stats as you add them. You can move to a larger plan from the dashboard.

Can n8n run AI agents on this server?expand_more

Yes. n8n's AI agent nodes send prompts to a model provider's API, such as OpenAI or Anthropic, and act on the answers, so the server needs no GPU. The VPS runs the workflows; the model runs at the provider. To run a model yourself, see our high-memory dedicated servers.

Is self-hosted n8n free?expand_more

The Community Edition is free to self-host under n8n's own licence. Check n8n's licence terms for what your use allows, especially if you build a product on it for others.

Will webhooks reach my server?expand_more

Yes. Every VPS has its own public IPv4 address, so services can call your webhooks directly. Point a domain at it, put HTTPS in front, and set n8n's WEBHOOK_URL to that address. The fixed address also suits services that only accept calls from allowlisted IPs.

What should I back up?expand_more

The n8n_data volume: it holds your workflows, credentials and the encryption key that protects them. Backups are an add-on on our plans, or copy the volume to a storage box on a schedule.

Running models yourself? High-memory dedicated servers hold large models in memory.

Guides for VPS

All guides

Server Locations