Basic
- 2 vCore CPU
- 2 GB Memory
- NVME 40 GB Disk
- Unlimited Traffic
AI and agents · Rio de Janeiro or São PauloOrder a Mega Dedicados Cloud VPS with Hermes Agent preinstalled, running on Debian 13 with Docker and HTTPS. You get the URL and an admin login, add your AI API key and start delegating tasks to the agent.


Hermes Agent requires 1536 MB of free memory, so the recommended plan is Basic. The more conversations you keep open at once, the more memory the agent uses, and that is where Plus and Pro give you more headroom.
Billed monthly in US dollars, with the app installed. Rio de Janeiro or São Paulo, same price. Plans start at the smallest size where the app runs well. Need more? Upgrade from the client area, without reinstalling.
With Rio de Janeiro selected, your VPS with Hermes Agent runs in Mega's original data center, on our own network with a presence at the Rio internet exchange (IX.br) and redundant transit. It is the lowest latency for users in Rio, Espírito Santo and Minas Gerais, and our support team works in the same city as the server.
With São Paulo selected, your VPS with Hermes Agent runs in our environment next to the busiest internet exchange in Latin America, on our own network with redundant transit. It is the shortest path to users across Brazil and South America and to banks, payment gateways and APIs hosted in São Paulo. Same price as Rio.
Hermes Agent is an open-source project under the MIT license, developed by Nous Research. It combines a language model with real tools, such as a terminal, files, a browser, web search, memory and task scheduling.
You interact with it through the web dashboard, which includes a terminal in the browser, or through external channels such as Telegram, Discord, Slack and WhatsApp. The agent keeps what it learned from one session to the next and supports ready-made skills.
On this server, Hermes runs inside a Docker container. The commands it triggers stay confined there and do not reach the VPS operating system. The AI model is up to you: OpenAI, Anthropic, OpenRouter, Gemini or any compatible service, using your own API key.
Source: Official Hermes Agent repository (Nous Research) on GitHub
The agent runs a command, reads the output and decides the next step on its own.
It visits sites, reads page content and returns an organized summary.
It writes scripts, runs tests and adjusts code inside the container.
Connect a Telegram bot and talk to the agent from anywhere.
Schedule reports and checks to run on set days and times.
Preferences and important information are stored for future conversations.


Images captured by Mega Dedicados on a test installation, with no customer data.
Terminal, files, code execution, browser, search and memory available to the agent.
Ready-made abilities, grouped by category, so you can start without writing anything.
Telegram, Discord, Slack, WhatsApp and more, with pairing to control who can talk to the agent.
Recurring tasks that run without you opening the dashboard.
MCP integration to add new tools to the agent.
Terminal chat, session history, model selection, keys and logs in one screen.
Select Basic, Plus or Pro with Hermes Agent and choose whether the server is in Rio de Janeiro or São Paulo, Brazil. The price is the same in both.
Pix, card or boleto, on a monthly, semiannual or annual cycle. There is no lock-in and you have 7 days to request a refund.
Once payment is confirmed, the VPS is created and Hermes Agent is installed with an HTTPS address and an admin account.
Your access details arrive by email and WhatsApp. On first login, add your provider and paste your AI API key.
Basic is the recommended plan and fits most cases. Move up a plan if you plan to open many terminal conversations at the same time.
| Plan | Best for |
|---|---|
| Basic 2 vCore · 2 GB | Recommended: meets Hermes' minimum and suits personal use with a few conversations open in the dashboard terminal. |
| Plus 2 vCore · 4 GB | For those who usually keep several terminal conversations open and want more memory and disk space. |
| Pro 4 vCore · 6 GB | For heavy or team use, with more cores and memory for parallel tasks and conversations. |
| Turbo 4 vCore · 12 GB | For those who delegate a lot, open many conversations at once or run other services alongside the agent. |
It is an open-source AI agent from Nous Research. Besides chatting, it runs commands, edits files, searches the web, keeps memory and runs scheduled tasks.
At Mega Dedicados, a VPS with the app installed starts at US$ 11.95 per month on the Basic plan, the recommended one. Hermes itself is free; AI model usage is billed by the provider of your API key.
The app requires 1536 MB of free memory. At idle it uses about 550 MB, and each conversation open in the dashboard terminal adds about 300 MB, so Basic is the starting point.
Yes. On first login, open the Models section, choose your provider (OpenAI, Anthropic, OpenRouter, Gemini or compatible) and paste the key. Without it, the agent has no model to answer with.
It runs in an isolated container, and the commands it executes stay inside that environment without reaching the VPS operating system. You still have root access via SSH to manage everything.
No. The dashboard requires a username and password, and channels like Telegram only respond to people you approve through pairing. Passwords are generated during installation and sent only to you.
The delivered version is fixed (0.21), so no update changes behavior without warning. You update whenever you want, using root access, preferably after testing.
Yes, each with its own address, as long as the total fits in the plan's memory. Since Hermes already uses a fair amount, consider Plus or Pro if you want to combine several.
Updated on October 19, 2026
Tell us what you want to run and for how many users; we will suggest the plan. The app arrives installed.