Host Hermes Agent on a VPS. Pre-installed, always on.
Run your open-source AI agent on a server that never sleeps. Pre-installed on Debian 13 with automatic HTTPS — connect your API key and start delegating.
Always-on automation
Bring any AI model
Full root access
Unmetered bandwidth
Run Hermes Agent from day one.
Every plan includes Hermes Agent installation, unmetered bandwidth, and full root access.
Length of Subscription
An agent that gets better the longer it runs.
An Agent That Works While You Sleep
Schedule daily briefings, recurring reports, and unattended jobs in plain language, delivered to Telegram, Discord, Slack, and more. Unattended work needs a server that never sleeps — a laptop isn't one.
Pre-Installed, Not a Weekend Project
The DIY route means install scripts, dependencies, and reverse-proxy configuration before you type a single prompt. Here, Hermes Agent and its web dashboard arrive running, served over automatic HTTPS.
An Agent That Learns How You Work
Hermes turns completed tasks into reusable skills, searches its own past sessions, and builds a picture of how you work. Keep copies of what it learns — sync its skill and memory files off-server as part of your routine.
Any Model. Your Keys. No Lock-In.
Hermes works with Nous Portal, OpenRouter, OpenAI, or any OpenAI-compatible endpoint. You bring your own API key and pay your provider directly — pick the model, control the spend, swap anytime.
Other VPS hosts meter your bandwidth. We don’t.
Most providers hand you a transfer allowance, then bill you per gigabyte the moment you exceed it — so a traffic spike becomes a surprise invoice. DreamHost doesn’t meter bandwidth at all.
is what 10 TB of monthly transfer costs on AWS Lightsail. On a DreamHost Self-Managed VPS, the same 10 TB costs $0.
| Provider | Bandwidth | Overage rate | Cost at 10 TB/mo | Cost at 50 TB/mo |
|---|---|---|---|---|
| DreamHost | Unmetered | No overage | $0 | $0 |
| AWS Lightsail | Metered | $0.09/GB | $810 | $4,410 |
| DigitalOcean | Metered | $0.01/GB | $90 | $490 |
| Vultr | Metered | $0.01/GB | $75 | $475 |
| Linode / Akamai | Metered | $0.005/GB | $45 | $245 |
No transfer caps, no overage fees, no surprise invoice when your traffic spikes — fair use applies.
Competitor figures are each provider's publicly listed overage rates as of June 2026, for outbound transfer beyond each provider's included transfer allowance (plan plus any account-level free-transfer grant). Providers publish these rates in USD; amounts shown in other currencies are approximate conversions. DreamHost Self-Managed VPS includes unmetered bandwidth with no per-GB or per-TB overage charges, subject to terms.
Your VPS. Run anything.
Choose your OS, deploy any app in minutes from your dashboard, and run whatever you build on infrastructure you fully own.
AI Agent
Open-source AI assistant for WhatsApp, Telegram, and Slack. Always-on private AI.
Automation & AI
Open-source AI agent that remembers your projects and gets more capable over time. Self-hosted, private, always on.
Workflow Automation
Connect apps and automate complex workflows without coding. 190k+ GitHub stars.
Development
Scalable JavaScript runtime for building fast network applications. Includes npm to manage thousands of open-source libraries.
Base OS
Universal default. Maximum compatibility with apps, tutorials, and resources.
Base OS
Community-governed. RHEL-compatible enterprise-grade Linux with a 10-year support lifecycle.
Base OS
Required for cPanel. RHEL-compatible enterprise-grade Linux for sysadmins.
Base OS
Lean, reliable, battle-tested. Preferred by experienced sysadmins.
Development
Run containers and compose multi-service stacks with Portainer CE for visual container management.
Database
Open-source Firebase alternative — database, auth, file storage, and real-time APIs.
Database
The most advanced open-source relational database. Runs under n8n, GitLab, and Supabase.
Productivity Suite
Files, calendars, contacts — your own private Google Drive on infrastructure you own.
Publishing
Modern platform for blogs, newsletters, and content monetization. Fast self-hosted publishing.
Development
Self-hosted Git, CI/CD pipelines, issue tracking, and a private container registry.
Development
Linux + Apache + MariaDB + PHP. 25 years of web history.
Development
Self-hosted PaaS to deploy and manage apps, databases, and SSL with a visual dashboard. Modern Heroku alternative.
Join thousands of businesses and developers worldwide on infrastructure that performs.
Launch your always-on Hermes Agent in minutes.
Frequently Asked Questions
- Two separate bills. The server is flat: Stack 4 starts at $8.99/mo for the first three months, then $15.99/mo, with Hermes Agent pre-installed and unmetered bandwidth. AI usage is separate — Hermes bundles no model, so you bring an API key from Nous Portal, OpenRouter, OpenAI, or any OpenAI-compatible endpoint and pay that provider directly, which means you pick the model and cap the spend.
- You don't have to. Pick the Hermes Agent image when you order your VPS and the server is provisioned in minutes — Hermes Agent and its web dashboard already running on Debian 13, served through a Caddy reverse proxy with automatic HTTPS. The DIY route — install script, dependencies, reverse-proxy configuration — is the part you're skipping.
- There aren't official ones — Nous Research publishes no RAM or disk minimums for Hermes Agent. The agent is a lightweight shell: models run at your provider's API, not on your server. Our guidance, not a Nous spec: start on Stack 4 and move up when you add gateways or heavier jobs — upgrading between Stack plans is straightforward.
- Any provider you choose. Hermes connects to Nous Portal, OpenRouter, OpenAI, or any OpenAI-compatible endpoint — your key, your provider, swappable anytime without rebuilding the server. Your VPS runs the agent, and inference happens at your provider's API — no model runs on the server itself, so a policy change at one provider never strands your agent.
- No, but you should be comfortable running a server. The hard parts of day one — install, reverse proxy, HTTPS — are done, and you talk to the agent in plain language through the web dashboard, or your messaging apps once you connect them. Day two is yours: expect SSH, updates, and patching. DreamHost supports hardware, network, and connectivity 24/7; Hermes itself is community-supported open source by Nous Research.
- Both are one-click images here, so it's preference, not lock-in. OpenClaw is known for channel breadth and its community-skill ecosystem; Hermes goes deep on learning — it turns finished tasks into reusable skills and searches its own past sessions. If you switch, Hermes ships a documented hermes claw migrate command that imports settings, memories, and skills from an OpenClaw install.
- Plan as if it won't. This image stores everything — skills, memory files, session history — on the server's root storage, and durability is not guaranteed. Make copies part of your routine: sync Hermes's skill and memory files off-server to git, object storage, or your own machine, so a rebuild costs you minutes instead of months of learning.
Need Some Help?
Whether you're stuck or just want some tips on where to start, hit up our experts anytime. We're here to help!
24/7/365 Through the Chat Widget
24/7/365 Through the Panel
24/7/365 Through the Panel