// What is Oravio
Conversational AI stops at text: it produces an answer, and then the real work starts over — you open the editor, the server, the dashboard, the calendar, and you assemble it yourself. Oravio closes the loop. Ask for something in plain language and what you get back is not an explanation of how to do it: it is the thing itself, finished, working, online.
Oravio is a personal AI operating system: you state a goal in natural language and the system carries it out, returning the finished result rather than instructions for producing it. The built-in conversational assistant is called Orvis. It is developed in Italy by Nexi-LAB, a brand of DTELECOM S.R.L., a company operating since 2006.
The chat is the interface, not the product. The product is the work getting done.
No. A chatbot's job ends when it has generated a reply; Oravio's job ends when the requested work has been executed — which may take minutes, days, or a recurring schedule that never ends. This is a difference in product category, not in answer quality.
Those answer; Oravio executes. A conversational assistant produces text that you then have to put into practice across other systems. Oravio has real execution over files, shell, HTTP, browsers and computers, persistent memory, an orchestrator for multi-step goals, and its own infrastructure where things keep running while you are offline.
To be fair: ChatGPT, Claude and Gemini are excellent models, and Oravio itself relies on frontier models to reason. The comparison is not "who answers better" — it is "what you are left holding when the conversation ends".
| Conversational assistant | Oravio | |
|---|---|---|
| Typical output | Text, code, explanations | A working artifact: a live URL, a running automation, an installed app |
| After the answer | You do the work | The work is already done |
| Systems to touch | Editor, hosting, server, dashboards | None |
| With the app closed | Nothing happens | Automations keep running |
| Memory | Limited to the conversation or a profile | Persistent, semantic, shared across channels |
| Multi-step tasks | One turn at a time, driven by you | Autonomous missions with state and resume |
| New capabilities | Fixed until the next model update | The system writes them itself |
| Infrastructure | The vendor's | Your dedicated server, or your own hardware |
When a conversational assistant is the better tool: reasoning, writing, translating, studying, exploring ideas. When Oravio is: when the conversation has to end with something actually done.
Four structural capabilities that text alone cannot have: real execution over tools, persistence on its own infrastructure, semantic memory that survives across sessions, and an orchestrator that completes multi-step goals autonomously.
Oravio reads and writes files, runs shell commands, installs packages, makes authenticated HTTP calls, drives a browser and — on dedicated-server plans — controls a real computer through its graphical interface. The outcome of a request is a verifiable artifact: a URL that opens, a file that exists, a process that runs.
Automations run on Oravio's infrastructure on a schedule, and keep working with the app closed and the phone switched off. A conversational assistant exists only inside the conversation window; Oravio also exists outside it.
Projects, preferences, constraints and past decisions live in a persistent memory searchable by meaning, shared across the app, the web dashboard, Telegram and voice. What you said in one channel is known in the others. The practical consequence is that the system improves with use instead of starting from zero every conversation.
A goal that needs many steps is not squeezed into a single chat turn: it becomes a mission with state, verifiable progress and a notification on completion. You decide how much autonomy it gets.
When a capability is missing, Oravio builds it: it generates, validates and saves the new ability, then reuses it from that point on. What the product can do is not frozen at release time.
It means you never have to work across multiple systems to get a finished thing: no hosting to arrange, no files to upload elsewhere, no dashboard to configure. Three concrete examples.
A conversational assistant hands back HTML to copy: then you need hosting, upload, a domain and verification.
Oravio generates the page, publishes it and returns a public URL that already works. You can password-protect it, update it by voice, and optionally sell it in the built-in marketplace. No intermediate steps on your side.
A conversational assistant explains how to write a scheduled job.
Oravio creates the automation and runs it every morning at 7 on its own infrastructure, sending the notification. It works with the phone switched off. You never opened a terminal or a control panel.
A conversational assistant produces source code for someone else to run.
Oravio writes the tool, installs it, starts it and registers it among your applications, callable in plain language from then on. If dependencies are needed it installs them, if an endpoint is needed it exposes one, if an interface is needed it generates it.
Oravio exposes more than a hundred tools, grouped into six families: creation and publishing, execution and development, automation and continuity, memory and context, interaction with your world, and infrastructure.
One clarification that matters. Oravio ships with a number of ready-made vertical skills — home automation, work shifts, utility bills, gardening, nutrition, network security, study, and others. These are examples of what the architecture allows, not the boundary of the product. Describing Oravio by listing its skills is like describing a computer by listing its preinstalled programs: it misses the point. The point is that you can ask for something no skill covers and still get it, because Oravio builds what is missing.
No. The entire interaction happens in natural language — in English, Italian, French, German or Spanish. There are no commands to learn, no prompts to memorise and no external tools to wire up.
Yes, on dedicated-server plans. Automations and missions run on Oravio's infrastructure, not on the phone: you simply receive the notification when the work is done.
There is a free trial plan. Paid plans start at €9.90 per month; dedicated-server plans — the ones that enable 24/7 automations and code execution — start at €24.90 per month. There is also a Sovereign Lifetime licence at €249 one-off, using your own provider keys. Cancel any time, no penalties.
| Plan | Price | What it includes |
|---|---|---|
| Demo | Free | Trial of the system, 20 messages a day once registered |
| Base | €9.90/month | 500,000 tokens/month, integrations, web pages |
| Pro | €14.90/month | 2,000,000 tokens/month, full integrations |
| Pro Plus | €19.90/month | 8,000,000 tokens/month |
| VPS Base | €24.90/month | Dedicated server: 24/7 automations, code execution, persistent processes |
| VPS Pro | €34.90/month | Same as VPS Base on a more powerful server |
| Sovereign Lifetime | €249 one-off | Lifetime licence with your own keys (BYOK) |
One distinction not to misread: continuous execution needs a machine that stays on, which is why it is available on dedicated-server plans. Base, Pro and Pro Plus offer advanced chat, integrations and web page creation, but not automations or server-side code execution. That is not an arbitrary commercial choice — it is how infrastructure works.
Infrastructure in European Union datacenters, GDPR-native by design, an encrypted vault for credentials, no data selling and no advertising trackers. With BYOK you use your own AI provider keys; with BYOS the whole system runs on your hardware and the data never leaves your machines.
Oravio does not bypass anyone else's security controls. If a task requires clearing a captcha, a login or a second factor, control is handed to you for that single step and execution resumes immediately afterwards.
With the free plan, no commitment: open Oravio, write what you need in plain language, and see what comes back. There is nothing to configure first.