NewOn-Demand Agents

AI agents that work for you, on your own server.

Hermes, n8n or OpenClaw installed on a LusoVPS server and ready to use. Connect your AI keys and the agent gets to work. Your data stays on your server.

  • Runs on your own server
  • Your AI keys, no markup
  • Dedicated resources
Examples of what each agent can do
  • Hermes working
    • Summary of today’s 14 messages sent on Telegram
    • New skill saved: weekly-report
    • Noted: you prefer reports on Mondays
  • n8n working
    • Invoice INV-2026-118 sent to accounting
    • Website request received, support ticket created
    • Today’s orders copied to the spreadsheet
  • OpenClaw working
    • Meeting with Anna booked for Thursday at 10:00
    • Replied on WhatsApp: the order ships tomorrow
    • Reminder sent: pay the rent by the 8th

Three agents, three ways of working

Pick the one that fits your work. Hermes and OpenClaw are open source; n8n is fair-code, with public source. The server is yours, so you can add another agent later.

The assistant that learns

Hermes Nous Research

An agent that keeps memory between conversations and writes its own skills as it works. Talk to it on Telegram, Discord, Slack, WhatsApp, by email or from the command line.

  • Persistent memory across conversations
  • Creates and reuses skills on its own
  • Splits large jobs across subagents
  • Works with OpenRouter, OpenAI or another compatible provider

Best for: anyone who wants a personal assistant that gets better with use.

Memory

  • Prefers short reports, on Mondays
  • Time zone: Lisbon
  • Invoices go to accounting@

Skills

  • summarise-messages
  • weekly-report new

Created after 4 similar requests

Illustrative example

Visual automation

n8n

Draw workflows in a visual editor and connect the services you already use: email, spreadsheets, CRM, invoicing, databases. Midway through a workflow, an AI agent reads, decides and summarises.

  • Visual editor, no coding needed
  • Hundreds of ready-made integrations
  • AI agent nodes inside your workflows
  • No workflow limit: only your server size limits you

Best for: businesses that want to automate repetitive work between apps.

Trigger New email
AI agent Sorts the request
Action Invoice → Accounting
Action Support → Ticket
Illustrative example

The assistant in your messaging apps

OpenClaw

A personal assistant that lives in the apps you already use. Message it on WhatsApp or Telegram and it takes care of the rest: calendar, reminders, research and replies.

  • WhatsApp, Telegram, Discord and Slack
  • Only answers the people you allow
  • Always on, 24 hours a day, on your server
  • Open project with a very active community

Best for: anyone who wants to hand off everyday tasks by message.

  1. Me · 09:12

    Book a call with Anna tomorrow at 10 and send her the agenda.

  2. OpenClaw · 09:12

    Done. Invite sent to Anna for tomorrow at 10:00, agenda attached.

  3. Me · 09:13

    Remind me 15 minutes before.

  4. OpenClaw · 09:13

    Will do. I’ll ping you at 9:45.

Illustrative example

How it works

  1. Pick the agent and the plan

    Hermes, n8n or OpenClaw, on a 2, 4 or 8 GB server.

  2. Get your server ready

    We deliver the server with the agent installed and configured.

  3. Connect your AI and start

    Add your OpenAI, Anthropic, OpenRouter or other provider key. From then on, the server is yours to manage.

Plans

The same price for any of the three agents. Monthly prices.

2 GB

To get started with one agent

€6.99/mo
  • 2 GB RAM
  • 2 vCPU
  • 20 GB disk
  • Hermes, n8n or OpenClaw pre-installed
  • Uses your own AI keys
Order

4 GB

For daily use and bigger workflows

€14.99/mo
  • 4 GB RAM
  • 2 vCPU
  • 40 GB disk
  • Hermes, n8n or OpenClaw pre-installed
  • Uses your own AI keys
Order

8 GB

For several agents or heavy workloads

€74.99/mo
  • 8 GB RAM
  • 4 vCPU
  • 100 GB disk
  • Hermes, n8n or OpenClaw pre-installed
  • Uses your own AI keys
Order

Prices exclude VAT. VAT at 23% is added where applicable.

Why LusoVPS

Your data, on your server
Conversations, memory and workflows stay on your server. Only the text the agent sends to the AI model leaves, to the provider you choose.
No markup on AI
Use your own keys and pay for usage directly to the model provider.
Dedicated resources
No overselling: the plan’s memory and processors are yours.
Real people in support
A Portuguese company, with an account manager you can contact directly.

Questions

Do I need an AI key?

Yes. The agent uses your OpenAI, Anthropic, OpenRouter or other compatible account, and you pay for usage directly to that provider.

Who manages the server after delivery?

You do. We deliver the server with the agent installed and configured; from then on you have full access to it and decide what to install.

Can I run more than one agent?

Yes. The server is yours: you can install another agent on it if the plan’s resources are enough, or order a second server.

Where is my data?

On your server. Conversations, memory and workflows do not go through our systems. The text of each request goes only to the AI provider you configure.

Can I change plans?

Yes. If your agent needs more memory or processing power, contact us to move to a bigger plan.

Put an agent to work today.

Pick a plan, get your server ready and connect your AI key.

Order from €6.99/mo