YOUR PROJECT + OUR INFRASTRUCTURE
Hermes Agent VPS
A VPS with Hermes Agent pre-installed · By Nous Research
An AI agent with memory, within reach of your chat apps.
Nous Research's AI agent with an HTTPS web dashboard, a login created at install and messaging channels such as Telegram, Discord and Slack.
- HTTPS web dashboard with the login created before the page opens
- Telegram, Discord and Slack over outbound connections, no open port
- You pick the provider and the model, and can change them any time
from R$59.90/mo
Pay with Pix, bank slip or card · You choose the location and the plan on the next screen. Nothing is charged until you confirm.
Ready in minutes
Automatic provisioning once the order is confirmed.
24/7 support
Our own team, by chat and ticket.
Pick the right plan for Hermes Agent
Every plan includes public IPv4 and IPv6, NVMe storage and a 1 Gbps uplink. Click one to reach the checkout with everything filled in.
How to size your project
The image is about 1 GB and the agent's tools (git, node, a browser) come and go on each turn, so the minimum is 2 GB and the installer creates swap below 4 GB. Since the models stay at the provider, load depends on what the agent runs. For browser tasks, start from 4 GB.
- Recommended for Hermes Agent
How to order Hermes Agent
Pick a plan
vCPU, memory and disk sized for your project. You can grow it later without reinstalling.
Confirm the order
Pix, bank slip or card. Nothing is charged until you confirm, and you have a 7-day guarantee.
Hermes Agent ready in minutes
The server comes up with Hermes Agent installed and configured. Access details appear in the panel.
What to use a Hermes Agent VPS for
Some of the most common uses. The server is yours — run whatever you like on it.
- Have an assistant that answers on Telegram, Discord or Slack, using the model of the provider you choose
- Talk to the agent from the terminal (hermes) when you are on SSH, without opening the dashboard
- Let the agent build up memories and skills between conversations, in a folder you back up yourself
- Try the same agent with different model providers, changing only the key and the model
Hermes Agent VPS specifications
- Virtualisation
- KVM, virtual server with your plan’s resources
- vCPU
- 2 to 48
- Memory
- 2 GB to 192 GB
- Storage
- NVMe, 30 GB to 2 TB
- Network
- 1 Gbps, public IPv4 and IPv6 included; DDoS protection available
- Datacenters
- Belo Horizonte (Brazil) · Dallas (USA)
- System
- Hermes Agent, pre-installed
- Management
- Web panel, API and MCP, VNC console, snapshots, backups and firewall
- Billing
- Monthly, quarterly, semi-annual or annual (discounted); Pix, bank slip or card
- Support and guarantee
- 24/7, 7-day money-back guarantee
About Hermes Agent
Hermes Agent, from Nous Research, is an AI agent that runs on your server, talks over the terminal or through messaging apps such as Telegram, Discord and Slack, and keeps sessions, memories and skills between conversations. It is for people who want an always-on assistant that uses the models of whichever provider they choose.
At Master da Web it runs from the official image (Hermes 0.21.5, version pinned), with the web dashboard behind Caddy over HTTPS. The dashboard starts with a login already in place: admin and a 24-character password generated at install, sent in the order e-mail and tested before the page is public. It is one shared login with no 2FA, which Hermes' own documentation describes as meant for trusted networks.
No model provider comes configured: under Keys you paste a key (OpenRouter, OpenAI, Anthropic, Nous Portal) and pick the model, or run hermes model over SSH. The messaging channels (Telegram or Discord bots) and the backup of /opt/hermes-agent/data/hermes, which holds keys and memories, are yours.
Software features
- Hermes 0.21.5 from Nous Research's official image, with an HTTPS web dashboard
- Admin login with a 24-character password, generated and tested before the page opens
- Messaging gateway (Telegram, Discord, Slack…) using outbound connections only
- Dashboard with Keys, Models and Channels; hermes model and hermes in the terminal
- Sessions, memories and skills kept in /opt/hermes-agent/data/hermes
- The gateway's API stays inside the container and is never published
- Dashboard login limited to 10 attempts per minute per address
- Version changes with hermes-agent-update, keeping your data
Requirements
Platform minimum
- Memory
- 2 GB
- Storage
- 30 GB
Below this the order is refused — the checkout raises the disk for you automatically.
Recommended for production
- vCPU
- 2 vCPU
- Memory
- 4 GB
- Storage
- 40 GB
What we would put it on, not a limit.
What your Hermes Agent gets at Master da Web
Manage infrastructure from one panel: security, recovery, connectivity and growth. Check prices and availability when configuring your server.
Snapshots
Take a snapshot of Hermes Agent before a risky update and roll back to it with one click, without reinstalling anything.
Automatic backups
Scheduled copies of the whole server, kept off the host. Restore the complete Hermes Agent whenever you need to.
Managed firewall
Rules by port and source, applied from the panel, no iptables editing. Open only what Hermes Agent needs.
Security Center
Continuous external scanning of your server, with findings prioritised and the fix spelled out for each one.
Private network (VPC)
Connect Hermes Agent to databases and other servers over an isolated network with nothing exposed to the internet. With a VPN Gateway for remote access.
DDoS protection
Volumetric attack mitigation for your Hermes Agent server. Check protection options and prices when ordering.
Grow without reinstalling
Add vCPU, memory and disk from the panel and pay only the difference, and Hermes Agent stays exactly as it is.
API and MCP for automation
Create, resize and manage servers through the public API — or straight from an AI assistant, over MCP.
Related applications
Langflow
Langflow
Visual editor for AI agents and flows served as APIs, with PostgreSQL and the superuser created at install, and public sign-up turned off.
from R$59.90/moView details →Open WebUI + Ollama
Open WebUI
A ChatGPT-style AI chat with models that run on the server itself: Open WebUI and Ollama installed on your VPS, one model ready to use.
from R$89.90/moView details →OpenClaw
OpenClaw
A personal AI assistant that gets things done over WhatsApp and Telegram, installed on your VPS with HTTPS and an e-mailed sign-in link.
from R$59.90/moView details →AI Coding Agents
AI Coding Agents
Claude Code, Codex CLI, Gemini CLI and OpenCode installed on a Debian 13 server, with Docker and SSH access for the dev user.
from R$89.90/moView details →
Know what to expect from day one.
Ready infrastructure, with the freedom to configure your application your way.
With your VPS
Your chosen image installed, the plan’s resources and a panel to manage the server. Administrative access is provided after provisioning.
Configure for your business
The order e-mail carries the dashboard address, the admin user and the password. Sign in, open Keys and paste your provider's key; then pick the model under Models. Messaging channels, firewall restrictions and the backup of /opt/hermes-agent/data/hermes are yours.
Expand when you need to
Backups, snapshot capacity, DDoS protection, private networking and other features are available as you need them. Check prices and availability when configuring your server.
You administer the system and application, including updates and data. For help with configuration or management, talk to our team about the scope of support.