Hermes agent · preinstalled · KVM VPS
An AI assistant that works for you, on your own server.
Hermes, the free open-source agent from Nous Research, remembers what you tell it, learns as it works and runs jobs on a schedule. We install it on a KVM VPS that is yours alone: add your AI account key, connect your chat apps, and it's ready. You look after the server yourself, or add paid Server Management at checkout and we'll do it.
Plans & pricing
Longer terms cost less per month. The price shown is per month at the term you pick, with the amount billed underneath.
Every plan includes Hermes agent preinstalled, comes up ready · Daily VM disk backup
H1 Starter
The smallest way in
Suggested for you₹1,604/mosave 17%
billed ₹19,248 per year
- 2 vCPU
- 2 GB RAM
- 80 GB NVMe SSD
- 2 TB bandwidth
- Note: 1 IPv4 + 1 IPv6
- Hermes agent preinstalled, comes up ready
- Note: Bring your own LLM key
- Note: Full root · KVM · self-managed
- Daily VM disk backup
H2 Recommended
Four times the memory of H1 Starter
Suggested for you₹2,622/mosave 17%
billed ₹31,459 per year
- 3 vCPU
- 8 GB RAM
- 120 GB NVMe SSD
- 3 TB bandwidth
- Note: 1 IPv4 + 1 IPv6
- Hermes agent preinstalled, comes up ready
- Note: Bring your own LLM key
- Note: Full root · KVM · self-managed
- Daily VM disk backup
H3 Pro
Double the memory of H2, for heavier work
Suggested for you₹3,835/mosave 17%
billed ₹46,020 per year
- 4 vCPU
- 16 GB RAM
- 160 GB NVMe SSD
- 4 TB bandwidth
- Note: 1 IPv4 + 1 IPv6
- Hermes agent preinstalled, comes up ready
- Note: Bring your own LLM key
- Note: Full root · KVM · self-managed
- Daily VM disk backup
- Seamless migration from your current host
- No lock-in, cancel anytime
- SLA-backed 99.9% uptime
- Plan prices exclude 18% GST, added at checkout for billing addresses in India
Each plan renews for the same term at the price in force then, shown before you pay. Promo codes apply to the first term only.
Bring your own LLM key
Hermes connects to whichever AI provider you choose: Anthropic, OpenAI, Google, OpenRouter, or a model you host yourself. You add your key after setup, so you pay that provider directly, at their rates, with nothing added by us. Model usage is billed by them rather than in your hosting price, and you can switch provider whenever you like. We help you wire it up and keep the key safe on the box.
What Hermes is
Hermes is Nous Research’s open-source AI agent: “the agent that grows with you”, in their words. It keeps its memory between conversations, turns what it learns into skills it writes itself, runs work on a schedule, and reaches you through the messaging apps you already use. It is MIT-licensed and free to run; the only thing you add is a key for the AI model you want it to think with. Since its first release in July 2025 it has passed 200,000 stars on GitHub (among the most-starred projects on the platform) and it is still moving quickly, with a new version most weeks.
The software is the open-source project from Nous Research · Hermes. We run the server it sits on; we are not affiliated with the project.
- 200,000+ GitHub stars
- 380+ Contributors
- MIT Open source, free
-
Memory that stays
It remembers across sessions: what you told it last month is still there. Its memory lives as plain files on your own server, not in someone else’s account.
-
Skills it writes itself
After it works through something involved, it can turn what it learned into a reusable skill, and sharpen that skill the next time it runs. You approve what it keeps.
-
Work that runs on time
Recurring jobs, described in plain English or as a cron expression. A gateway process checks every minute, so things fire whether or not you are at your desk.
-
Reachable where you already are
Telegram, WhatsApp, Slack, Discord, Signal, email and twenty-odd more: one agent, answering from whichever you happen to open.
-
Any model you like
Over forty providers: Nous Portal, OpenRouter, OpenAI, or a model you run yourself. Nothing is bundled and nothing is locked; the key stays yours.
Why it wants a server
The scheduler only runs while the gateway is running, so a laptop that sleeps, closes or hops networks takes the agent down with it; Nous point at a small always-on machine instead. On your own instance it stays awake, keeps its memory and skills on disk you control, and answers from Telegram while you are nowhere near a desk.
Why run Hermes with us
Private
Private to you
Hermes runs on a server that's yours, in the region you choose. Its memory, skills and connected accounts stay on your server, not a shared platform.
Contained
Kept apart from everything else
Hermes' makers advise running it apart from other systems. Your copy starts on its own virtual machine, used by you alone, not a slice of a shared platform, and you have full control to lock it down further.
One-click
Ready when you order
The server arrives with Hermes installed and set to start by itself, ready for your AI key and your chat apps. Nothing technical to set up before your first conversation.
Yours
You're in charge, with help on tap
Hermes releases updates about once a week. On a self-managed server you decide when to update, with a best-effort full backup of the server every day as a safety net. Paid Server Management is at checkout if you'd rather hand the server side to us.
How it works
-
01
Pick a plan
Three tiers on KVM hardware: start where your workload fits and move up later.
-
02
The image does the setup
Pick where the server lives (India, Germany or the US) and your VPS deploys with Hermes installed and registered as a service. Optional extras (offsite backups, paid Server Management) are there to take or leave.
-
03
Add your model key and channels
Point Hermes at your provider: Nous Portal, OpenRouter, OpenAI, or a model you host (it wants a 64k+ context window). Inference is billed by them, and never appears on your hosting bill.
-
04
Yours to run
A self-managed VPS with full root: you decide when to update Hermes and what runs beside it. Hardware, network and the virtualisation layer are on us; Server Management tiers are at checkout if you want us on the box.
ssh root@your-vpshermes ██╗ ██╗███████╗██████╗ ███╗ ███╗███████╗███████╗ █████╗ ██████╗ ███████╗███╗ ██╗████████╗██║ ██║██╔════╝██╔══██╗████╗ ████║██╔════╝██╔════╝ ██╔══██╗██╔════╝ ██╔════╝████╗ ██║╚══██╔══╝███████║█████╗ ██████╔╝██╔████╔██║█████╗ ███████╗█████╗███████║██║ ███╗█████╗ ██╔██╗ ██║ ██║██╔══██║██╔══╝ ██╔══██╗██║╚██╔╝██║██╔══╝ ╚════██║╚════╝██╔══██║██║ ██║██╔══╝ ██║╚██╗██║ ██║██║ ██║███████╗██║ ██║██║ ╚═╝ ██║███████╗███████║ ██║ ██║╚██████╔╝███████╗██║ ╚████║ ██║╚═╝ ╚═╝╚══════╝╚═╝ ╚═╝╚═╝ ╚═╝╚══════╝╚══════╝ ╚═╝ ╚═╝ ╚═════╝ ╚══════╝╚═╝ ╚═══╝ ╚═╝ ╭──────────────────────────────── Hermes Agent v0.15.1 (2026.5.29) · upstream 689ef5e2 ────────────────────────────────╮│ Available Tools ││ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢀⣀⡀⠀⣀⣀⠀⢀⣀⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ browser: browser_back, browser_click, ... ││ ⠀⠀⠀⠀⠀⠀⢀⣠⣴⣾⣿⣿⣇⠸⣿⣿⠇⣸⣿⣿⣷⣦⣄⡀⠀⠀⠀⠀⠀⠀ browser-cdp: browser_cdp, browser_dialog ││ ⠀⢀⣠⣴⣶⠿⠋⣩⡿⣿⡿⠻⣿⡇⢠⡄⢸⣿⠟⢿⣿⢿⣍⠙⠿⣶⣦⣄⡀⠀ clarify: clarify ││ ⠀⠀⠉⠉⠁⠶⠟⠋⠀⠉⠀⢀⣈⣁⡈⢁⣈⣁⡀⠀⠉⠀⠙⠻⠶⠈⠉⠉⠀⠀ code_execution: execute_code ││ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣴⣿⡿⠛⢁⡈⠛⢿⣿⣦⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ computer_use: computer_use ││ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠿⣿⣦⣤⣈⠁⢠⣴⣿⠿⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ cronjob: cronjob ││ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠈⠉⠻⢿⣿⣦⡉⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ delegation: delegate_task ││ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠘⢷⣦⣈⠛⠃⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ discord: discord ││ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢠⣴⠦⠈⠙⠿⣦⡄⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ (and 22 more toolsets...) ││ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠸⣿⣤⡈⠁⢤⣿⠇⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ││ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠉⠛⠷⠄⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ Available Skills ││ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢀⣀⠑⢶⣄⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ autonomous-ai-agents: claude-code, codex, hermes-agent, kanban-codex-... ││ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣿⠁⢰⡆⠈⡿⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ creative: architecture-diagram, ascii-art, ascii-video, b... ││ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠈⠳⠈⣡⠞⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ data-science: jupyter-live-kernel ││ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠈⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ devops: kanban-orchestrator, kanban-worker, webhook-sub... ││ email: himalaya ││ claude-opus-4.6 · Nous Research gaming: minecraft-modpack-server, pokemon-player ││ /root general: dogfood, yuanbao ││ Session: 20260807_085656_e9a25f github: codebase-inspection, github-auth, github-code-r... ││ mcp: native-mcp ││ media: gif-search, heartmula, songsee, spotify, youtub... ││ mlops: audiocraft-audio-generation, dspy, evaluating-l... ││ note-taking: obsidian ││ productivity: airtable, google-workspace, linear, maps, nano-... ││ red-teaming: godmode ││ research: arxiv, blogwatcher, llm-wiki, polymarket, resea... ││ smart-home: openhue ││ social-media: xurl ││ software-development: debugging-hermes-tui-commands, hermes-agent-ski... ││ ││ 17 tools · 85 skills · /help for commands │╰──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯ Welcome to Hermes Agent! Type your message or /help for commands.Questions
Which AI models does Hermes use?
Your own, via Nous Portal, OpenRouter, OpenAI, or a model you host; there are more than forty providers to pick from. Hermes ships no bundled model; you add your key and pick the provider. That inference is billed by them, separately from your hosting. One real constraint: Hermes needs a model with at least a 64,000-token context window, so the smallest and oldest models are out. Check that before you pick, or ask an engineer if you are unsure.
Which plan should I start on?
H2 Recommended is the tier we recommend, as its name says. H1 Starter is the smallest way in, and H3 Pro doubles the memory of H2 for heavier work. Whichever you pick, you can move up later, and an engineer can help you size it if you tell us what the agent will do.
Do I need to install anything?
Not to get it running. Ordering provisions a VPS with Hermes pre-installed and registered as a service, so it comes back up on its own after a reboot. Connecting your model key and channels is yours to do; Hermes’ onboarding walks you through it, and an engineer can point you right if you get stuck.
What does “self-managed” mean here?
The base plans are a real KVM VPS with full root that you run: OS updates, Hermes updates and housekeeping are in your hands. Hardware, network and the virtualisation layer are ours. If you want our engineers on the box (updates, monitoring, hardening), add a paid Server Management tier at checkout.
Are my data and key safe?
Everything runs on your own single-tenant instance, in the data centre region you choose at checkout, and the model key lives on your box, held by you. Hermes keeps its memory and skills as plain files on that disk, so you can read what it remembers and take a copy whenever you like.
Are backups included?
Yes, at VM level: we take a daily full-disk backup of your instance on a best-effort basis, and it is what we use to bring the whole server back if something goes badly wrong. Restoring an individual file or database is yours to do on a self-managed box, so keep your own copies of anything you cannot lose; the “Offsite Backups for Data” option at checkout adds offsite copies.
Is the price final?
It is the full hosting price for the term you choose. Promotional discounts apply to your first term. If your billing address is in India, 18% GST is added at checkout.
Hermes or OpenClaw?
Both are open-source AI assistants that live on a server of your own and answer through your chat apps, on the same self-managed KVM line. Hermes is built around learning: it turns what it works through into skills it writes and refines itself. OpenClaw is built around one gateway for around two dozen chat channels, with voice, a live canvas and companion apps. If you are unsure which suits what you have in mind, ask an engineer.
Ready to give Hermes a home?
Order and it comes up ready for your key and your channels. Unsure which tier, or whether your model choice fits? Ask an engineer; you'll get a straight answer.