Hermes Agent hosting
Managed Hermes Agent Hosting: Your Own Agent, Running in About a Minute
Updated 10 October 2026
Hermes Agent is open-source and free to install. The catch is that it only becomes useful once it runs all the time: answering in Telegram while your laptop is shut, firing a scheduled job at 8am, remembering last week. That means a machine that stays on, and someone looking after it.
Managed Hermes Agent hosting is the version where that someone is not you. This page explains what “managed” should actually cover, gives you a checklist for judging any provider (including us), and then lays out exactly what ReadyHermes handles and what stays yours.
What “managed” should mean for Hermes Agent
The word gets stretched. A cloud server with a one-click install is not managed. You still patch it, restart it and pay for whatever it does at 3am. For an always-on agent, we think “managed” should answer six questions with “not you”.
- Who runs the server? You should never log in to a machine to keep the agent alive.
- Who handles updates? Hermes Agent changes often. Operating system patches and agent upgrades should arrive without a weekend of your time.
- Who owns the model keys? Either the host includes AI so you need no key, or you can bring your own, or both. Hunting for API keys on day one is where a lot of people stall.
- Who watches uptime? If the agent falls over at night, something other than you should notice.
- What about backups? You should be able to get your agent’s data out, as a file, on demand.
- Who wires up channels? Telegram, WhatsApp and Discord each have their own setup steps. A managed host should make them a form, not a project.
If you want the full trade-off against running it yourself, we wrote that up in self-hosting vs managed hosting.
A checklist for judging any managed host
Use this on every provider you look at, ReadyHermes included. Most of these take two minutes to check on a pricing page or in a trial.
Is the agent really yours?
Ask whether you get a dedicated agent with its own memory and skills, or a seat in a shared chatbot. Hermes Agent’s value is that it learns you. A shared instance cannot do that safely.
What is the real monthly cost?
Look past the introductory price. Check the term length, the renewal price and whether model usage is included or billed separately. A cheap plan with a 24-month lock-in is a different product from a monthly one.
Can you leave with your data?
Look for a backup you can download and a way to delete your account yourself. If the only exit is “email support”, treat that as a warning.
Do you get access, or only a chat box?
A good host lets you reach the agent’s files and, if you want, a shell or an API. A closed box is fine for some people, but know which one you are buying.
Is the free test honest?
You should be able to try the real thing before paying, and the page should say what happens when the trial ends. Vague trial terms usually hide a card requirement or an automatic charge.
Are the limits stated?
Memory, storage and AI usage all have ceilings somewhere. Hosts that publish them are easier to trust than hosts that say “unlimited”.
One fair point against us: on raw specs, a VPS usually gives you more RAM and disk per dollar. If you want to tinker with the machine itself, a VPS is the better buy. Our own comparison is in ReadyHermes vs Hostinger.
What ReadyHermes manages, and what stays yours
ReadyHermes is an independent hosting service. We are not affiliated with Nous Research, the team behind Hermes Agent. Each account gets its own dedicated Hermes Agent, with its own memory and skills, isolated from other users. It comes with long-term memory, skills it writes for itself, scheduled jobs, web search, a real web browser and its own Linux computer.
| What we handle | What you control |
|---|---|
| The server and Docker setup. You never see either. | What the agent does: your prompts, schedules and skills. |
| Updates to the agent and the machine under it. | When to start, stop or restart it (Bot tools, Power). |
| 24/7 monitoring. | Whether to use the included AI or your own model key. |
| Included AI (DeepSeek V4.1 Flash), no key needed. | Your own key from OpenRouter, OpenAI, Anthropic, Google Gemini, DeepSeek and others, on paid plans. |
| Storing your account and agent data in the EU. | Downloading a backup .zip, restoring it, and browsing or changing files. |
| Channel setup screens for Telegram, WhatsApp, Discord, WeChat and Feishu/Lark. | Which chat apps you connect, and which apps (Gmail, Notion, GitHub) you sign in to. |
| Billing through Waffo, the merchant of record. | Your plan, your card, cancelling, and deleting your account. |
The Bot tools, in plain terms
“Managed” should not mean “locked”. Inside the dashboard, Bot tools give you the controls that usually come with owning a server.
- SSH. Paste a one-time auth key from your own Tailscale account and your agent joins your private network. No public port is opened. You connect with a command like
ssh root@<its name>, which Bot tools shows you. - API. Switch on an OpenAI-compatible endpoint with a key, for your own scripts and servers. It is server-to-server, so it is not meant for browser web apps.
- Backup and Restore. Download your agent as a .zip and load it back later. Managed settings, such as the AI model and connected chat apps, are not restored from a backup.
- Files. Browse, upload, download, rename and delete. There is no in-browser editor, so for editing use SSH.
- Desktop and My computer. Watch the agent’s desktop live, or give it limited access to folders on your Mac, Windows or Linux machine. It is read-only by default, and saving files or running commands are separate permissions that stay off until you turn them on.
On privacy: to write a reply, your messages go to the AI model provider, which may be outside the EU. We do not train AI on your chats and we do not sell data.
The limits, stated up front
- Lite has 2 GB of memory and 10 GB of storage. That is enough for chat, scheduled jobs and light browsing. Heavy work wants Pro.
- AI usage is included on both plans, with fair-use limits. Pro is a bigger machine, not a bigger AI allowance.
- The free trial lasts 30 minutes from the first time the agent runs. Then it pauses. You have 24 hours to choose a plan; otherwise it is deleted.
- WhatsApp uses the linked-device session, so the phone that hosts the bot must stay online. Upstream Hermes recommends a dedicated number for it.
- If you cancel, the agent works to the end of the paid period, stops, and is deleted 7 days later unless you subscribe again.
Pricing
| Plan | Price | Resources |
|---|---|---|
| Free trial | $0, no card | 30 minutes of Lite |
| Lite | $5/month | 1 vCPU, 2 GB memory, 10 GB storage |
| Pro | $10/month | 2 vCPU, 4 GB memory, 40 GB storage |
Billing is monthly and you can cancel any time. Prices exclude local tax where it applies. Pro is the same agent resized in place, so you can switch whenever you like and lose nothing. Payments go through Waffo, and its customer portal handles plan changes, card updates, invoices and cancellation. More on the budget end in cheap Hermes Agent hosting, or see pricing on the home page.
How to get started
- Go to readyhermes.com and click “Start free trial”. Sign up with email and password, or continue with Google.
- Wait on the setup screen while the agent starts. It takes about a minute.
- The dashboard opens on Assistant, a chat box with starter prompts. Ask for something real, like a weekday 8am briefing.
- Under Channels, connect Telegram or another chat app. Scheduled results are delivered to a connected chat app, so do this if you want briefings and reminders to reach you.
The full walkthrough, with every channel and Bot tool, is in the step-by-step guide. When you are ready, start the free trial.
Questions
What does managed Hermes Agent hosting include?
On ReadyHermes it includes your own dedicated Hermes Agent, the server it runs on, updates, 24/7 monitoring, included AI (DeepSeek V4.1 Flash) and setup screens for chat channels. You do not need a server, Docker or an API key.
Do I need my own API key?
No. AI usage is included on both plans, with fair-use limits. On paid plans you can paste your own key from providers such as OpenRouter, OpenAI, Anthropic or Google Gemini, and your provider then bills you directly for that usage. You can switch back to the included AI at any time.
Can I still get shell access on a managed plan?
Yes. Bot tools has an SSH tab. You paste a one-time auth key from your own Tailscale account, your agent joins your private network, and no public port is opened. There is also an OpenAI-compatible API and a file manager.
Can I take my agent with me if I leave?
While your trial or plan is active, you can download your agent as a .zip from Bot tools (tick Conversations, Workspace files and memory, and Skills to include them), so download it before access ends. You can also download a copy of your data or delete your account from Privacy & data. Managed settings such as the AI model and connected chat apps are not restored from a backup.
What happens when the free trial ends?
The trial is 30 minutes of Lite and needs no card. The clock starts when the agent first runs. When it ends the agent pauses, and you have 24 hours to choose a plan to keep it with its memory and files. Otherwise it is deleted.
Is ReadyHermes affiliated with Nous Research?
No. Hermes Agent is the open-source agent made by Nous Research. ReadyHermes is an independent hosting service that runs it for you.
How much does it cost?
Lite is $5 per month with 1 vCPU, 2 GB memory and 10 GB storage. Pro is $10 per month with 2 vCPU, 4 GB memory and 40 GB storage. Billing is monthly, you can cancel any time, and prices exclude local tax where it applies.