Persistent memory and skills
Sessions, memory, learned skills, and scheduled-job state stay on the server's NVMe storage.
Hermes Agent VPS hosting
Run the open-source Nous Research agent on a persistent Linux server. Its gateway, memory, skills, scheduled tasks, and messaging connections stay on the server.
$3.89/mo
- <30s
- Linux image deploy
- NVMe
- Every plan
- up to 25 Gbps
- Network
ubuntu@server · sshonline
$ curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
Hermes Agent installed successfully
$ hermes setup --portal
Provider and Tool Gateway configured
$ sudo hermes gateway install --system
Gateway service installed · active
$
- Memory
- Persistent
- Skills
- Reusable
- Scheduler
- Gateway
- Runtime
- Linux service
Hermes Agent VPS plans
Choose a server size for your Hermes Agent workload.
- Linux images
- NVMe storage
- DDoS protection
- Network capacity up to 25 Gbps
Choose a size
Model-provider usage is separate from the VPS price. See the complete plan list for additional sizes.
Hermes Agent system requirements
| Profile | Hermes guidance | Arct Cloud match |
|---|---|---|
| Official minimum | 1 core · 1 GB RAM · 500 MB data volume | cvm.pico · 1 vCPU · 1 GB · 15 GB NVMe · $3.89/mo |
| Official recommended | 2 cores · 2–4 GB RAM · 2+ GB data volume | cvm.micro · 2 vCPU · 4 GB · 40 GB NVMe · $8.89/mo |
| Browser automation | At least 2 GB RAM for Playwright and Chromium | cvm.micro · 2 vCPU · 4 GB · 40 GB NVMe · $8.89/mo |
What runs on the VPS
The server hosts the Hermes Agent runtime and state. Inference can stay with the model provider you configure.
Gateway as a system service
Install the Hermes gateway as a Linux service so messaging and scheduled tasks resume after boot.
Messaging connections
Connect Telegram, Discord, Slack, WhatsApp, Teams, and other supported platforms.
Provider choice
Configure Nous Portal, OpenRouter, OpenAI, or another provider supported by Hermes Agent.
Scheduled work
The gateway runs one-time and recurring jobs through Hermes Agent's built-in scheduler.
A separate tool environment
Keep agent files, dependencies, and task credentials on a Linux server with root SSH access.
One server, three parts
Messages reach the gateway, Hermes uses its stored context and tools, then the configured provider handles inference.
ChannelsTelegram · Discord · SlackMessages and scheduled results
Your Arct Cloud VPSHermes gatewayMemory · skills · cron · tools
Configured servicesModel provider · APIsSeparate accounts and usage
Install Hermes Agent in four steps
Deploy Linux
Choose a plan and current location, then deploy a Linux image.
Install Hermes
Connect with root SSH and run the official Nous Research installer.
Configure a provider
Run the setup flow for Nous Portal or another supported model provider.
Start the gateway
Configure messaging, then install the gateway as a Linux system service.
Official Linux install path
Start from a clean Linux image. The installer and setup commands remain owned by the Hermes Agent project.
$ curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash$ hermes setup --portal$ sudo hermes gateway install --systemHermes Agent VPS questions
Hermes Agent is an open-source AI agent built by Nous Research. It keeps memory across sessions, creates reusable skills, runs scheduled tasks, uses tools, and connects to messaging platforms.
Read the Hermes Agent documentationNo. Arct Cloud provides the Linux VPS. Install Hermes Agent with the official Nous Research installer, then configure your model provider, tools, and messaging gateway.
The official Docker guide lists 1 core and 1 GB RAM as the minimum, and 2 cores with 2–4 GB RAM as recommended. cvm.pico matches the minimum; cvm.micro matches the 2-core, 4 GB profile.
Check the official resource guideNo. The VPS runs Hermes Agent, its gateway, memory, skills, and tools. Model access through Nous Portal, OpenRouter, OpenAI, or another supported provider is configured and billed separately.
Yes. The official Docker guide says browser automation is the most memory-intensive feature and calls for at least 2 GB RAM. A 2 vCPU, 4 GB cvm.micro plan leaves more room for Chromium and the gateway.
Yes. Hermes Agent has a messaging gateway for Telegram, Discord, Slack, WhatsApp, Teams, and other supported platforms. The gateway can run as a Linux system service at boot.
Only open the ports your setup needs. Hermes uses port 8642 for its optional API server; the official guide requires an API key when it is exposed beyond localhost.
Explore Arct Cloud
Give Hermes Agent a persistent Linux server
Start at $3.89/mo for the official minimum profile, or choose the 2 vCPU and 4 GB cvm.micro plan at $8.89/mo.