Client area
Deploy Hermes Agent on VPS in one click
Run a self-hosted Hermes Agent on VPS that stays online around the clock. Pick the preinstalled image at checkout, connect to Telegram, Slack, or email, and point it at a free model with $0 in API fees.
What can you do with self-hosted Hermes Agent on VPS
Hermes is an open-source agent from Nous Research. On your own VPS hosting, it keeps context between chats, runs subagents, drives a real browser, and writes its own reusable skills. There are some Hermes Agent use cases.
One agent, living in your chat apps
Reach it from Telegram, Slack, Signal, Teams, and 20+ other channels. Every channel talks to the same agent with the same memory, and it keeps running after you log off.
Bring your own model, switch when you want
Connect OpenRouter, Nous Portal, OpenAI, NVIDIA NIM, or your own endpoint. Moving between them takes one line and a restart, avoiding vendor lock-in.
Keep a local model when the data shouldn't leave
Point Hermes at Ollama or LM Studio on the same server, and prompts never reach an external API. You pay nothing per token, and no third party stores your inputs.
Subagents for parallel work
Hand research, scraping, and code to separate subagents, each with its own terminal. They run at the same time rather than waiting in a single queue.
A browser for tools without an API
Many dashboards never shipped an API. Hermes drives a real browser through them, clicking, filling forms, taking screenshots, and extracting data.
Memory and skills that build up over time
Solve a problem once, and Hermes saves it as a reusable skill. Persistent memory means that month two already handles what you explained in month one.
Choose a VPS plan for Hermes Agent
A self-hosted agent runs nonstop, so the only spec that bites is RAM. Start at Start VPS to try it; move to Medium VPS if it's your daily driver. Every plan includes one dedicated IPv4 address, free weekly backups, KVM, and full root/SSH access. RAM, drive, and other specs are scalable later.
- 40 CountriesFrom$10.19 / 1 month$9.68 with is*smart
- Xeon 2x2.20 GHz CPU
- 2 Gb RAM
- 30GB SSD Drive
- 3 Tb Bandwidth
- 40 CountriesFrom$21.24 / 1 month$20.18 with is*smart
- Xeon 3x2.20 GHz CPU
- 4 Gb RAM
- 40GB SSD Drive
- Unmetered Bandwidth
- 40 CountriesFrom$33.99 / 1 month$32.29 with is*smart
- Xeon 4x2.20 GHz CPU
- 8 Gb RAM
- 50GB SSD Drive
- Unmetered Bandwidth
Boost Your VPS Server Resources
Check out how you can change the configuration of your VPS. You can purchase additional drive space, increase CPU performance, add RAM, and take advantage of additional services without having to switch to a new plan.
How to deploy Hermes Agent on a VPS
The preinstalled image does the heavy lifting: Python, Node.js, and Hermes are already there. Full walkthrough with every edge case lives in our step-by-step Hermes Agent setup guide.
Here, we'll describe the process for Telegram, but when setting it up, you can choose any other available messaging app or email service.
Step 01. Order a VPS with the Hermes Agent image
In the configurator, choose Ubuntu 24 and select the Hermes Agent image under OS. After payment, you get an IP and root password by email. Connect over SSH.
ssh root@YOUR_VPS_IPStep 02. Pick a provider and a free model
On first login, the setup wizard launches automatically. If it doesn't, run hermes.
In the wizard, choose OpenRouter, paste an API key from openrouter.ai/keys, and select a model marked free. A working one at the time of writing:
nvidia/nemotron-3-super-120b-a12b:freeStep 03. Connect Telegram and run as a service
Create a bot with @BotFather, copy the token, and paste it when the wizard asks. Lock the bot to your own account with your numeric ID (message @userinfobot to get it). When prompted, run the gateway as a system service so it auto-starts on reboot.
Step 04. Test it
Open Telegram, find your bot, and send hi. If it replies, you're done.
Common issues
Field-tested gotchas.
Bot runs but won't answer — “The model provider failed after retries.” The free model rotated out of OpenRouter (it happens). Switch it and restart:
hermes config set model.default nvidia/nemotron-3-super-120b-a12b:freesudo systemctl restart hermes-gatewayThe bot ignores you entirely. The gatekeeper is TELEGRAM_ALLOWED_USERS in /root/.hermes/.env, not config.yaml. If it's set, anything in config.yaml is ignored. Add your numeric ID there and restart.
A manual edit “saved,” but nothing changed. Hermes parses config.yaml quietly and falls back when a section is malformed, including from invisible characters pasted into nano. After any hand-edit, run hermes config show to confirm it actually loaded what you wrote.
Authentication silently fails. In config.yaml, api_key_env is the name of the environment variable, not the key itself. The actual key lives in /root/.hermes/.env. Paste a raw sk-… into api_key_env and Hermes won't error.
The full set, including auxiliary-model cascades and the Tirith scanner, is in the Hermes troubleshooting guide.
is*hosting infrastructure is a solid base for Hermes Agent
Your cores, your RAM via KVM
KVM virtualization with dedicated resources means what you pay for is what runs. There is no shared pool deciding when your gateway gets OOM-killed.
No middlemen. Full control
We own our hardware and run Hi-End equipment across 40+ locations. Root on every plan, and resources you can scale when your memory and skill library grow.
Backed by Tier 3+ data centers
Every location meets Tier 3 standards or higher — redundant power and network, plus physical security — so the server hosting your agent keeps running.
Frequently asked questions about Hermes Agent hosting
What is Hermes Agent VPS hosting?
It's running the open-source Hermes Agent by Nous Research on a virtual private server instead of your laptop. The agent stays online 24/7, connects to messaging apps, and works with dozens of LLM providers. On is*hosting you deploy it from a preinstalled image.
Why host Hermes Agent on VPS instead of running it locally?
A local agent stops the moment you close the lid. On a VPS, it runs around the clock, keeps a stable IP, and handles scheduled tasks and subagents without tying up your machine. Self-hosting Hermes Agent on a VPS also keeps your data and keys on hardware you control.
Do I need a GPU for Hermes Agent VPS hosting?
No GPU for a standard setup, as the model runs on a provider's API. The Start plan (2 GB RAM) is the floor; Medium (4 GB) is comfortable for subagents to run Hermes Agent on is*hosting. A GPU only matters if you run a local LLM with Ollama.
Which messaging platforms does Hermes Agent support?
More than 20. Telegram, Discord, Slack, and WhatsApp are the common ones, but the setup wizard also covers Signal, Matrix, Microsoft Teams, Google Chat, iMessage (via BlueBubbles), LINE, IRC, Email, SMS through Twilio, and several China-market apps like WeChat, WeCom, Feishu/Lark, and DingTalk. You tick the ones you want in the wizard's “Select platforms to configure” step.
Can Hermes Agent connect to OpenAI, OpenRouter, or custom endpoints?
Yes. OpenRouter, Nous Portal, OpenAI, NVIDIA NIM, or your own endpoint. Switch with one command: hermes config set model.default MODEL. Free OpenRouter models cost $0; a paid model like Claude Haiku 4.5 runs a few dollars a month for typical personal use. Self-hosting Hermes Agent on VPS means you can choose the model provider.
What is the difference between Hermes Agent and OpenClaw?
Both are self-hosted AI agents. Hermes leans toward a messaging-first personal assistant with subagents and a growing skill library; OpenClaw is more of a general assistant framework. If you're weighing them, see our OpenClaw on VPS page.
More from is*hosting
VPS
Deploy VPS instances across 40+ locations. Manage multiple servers from one account — useful for running additional subagents or a separate local-model node.
Explore
From $5.94 /mo
Dedicated server
When a VPS isn't enough — dedicated hardware for high-traffic production, GPU workloads (local LLMs with Ollama), or large databases.
Explore
From $75.00 /mo