An Aivell is an on-premise AI colleague: the model runs on a small box inside your office, behind your firewall. Documents, emails and data are processed onboard — nothing is uploaded, nothing leaves your network. Unplug the internet and it keeps working. Privacy by design, not by promise.
Every “private AI” pitch ends the same way: a quiet line in the terms about processing on external servers. An Aivell has no such line. The models that read your documents, draft your replies and work through your chores run on a small box on your shelf. Inference happens onboard, behind your firewall, on hardware you own.
That changes what “private” means. Not “encrypted on the way to someone else’s data center”. Not “we promise not to train on it”. Private the way your filing cabinet is private: the contract, the CV, the payroll sheet never go anywhere. They are read, understood and filed by a machine that sits two rooms away from the people they concern.
There is a simple way to audit a privacy claim: pull the network cable. Do that to most AI appliances and they fall silent, because the thinking was always happening somewhere else. Do it to an Aivell and it keeps working — reading scans and photos, updating spreadsheets, answering questions from the knowledge base, filing documents on the network share. Steps that genuinely need the outside world, like fetching a portal page or sending an email, wait for the connection to return. Everything inside your walls carries on.
That is not a party trick; it is the architecture. Air-gapped capable by design, cloud not required, no exceptions buried in a settings page.
If you handle client files, patient records or payroll, “we send it to a cloud AI” is a sentence you would rather not put in writing. On-premise inference dissolves the problem instead of managing it: under GDPR Article 25, keeping inference local is privacy by design — the data never reaches a processor, so there is nothing to assess, transfer or apologize for. NIS2 and the AI Act push in the same direction: know where your data goes, minimize who touches it. “Nowhere” and “no one” are easy answers to defend.
Inside the box, data is encrypted at rest and every action lands in the audit trail. Outside the box, there is nothing to find. When a client, an auditor or a works council asks where the AI sends their data, you point at the shelf.
No. The models that read, reason and write all run on the box. There is no cloud fallback and no telemetry carrying your text away. When a chore does need the internet — fetching a supplier portal, sending an email — that traffic is yours to see and yours to firewall.
Your Aivell keeps working. Inference is onboard, so chores that touch local systems — network folders, your database, the knowledge base — run exactly as before. Steps that wait on the outside world resume when the connection does.
Yes. GDPR Article 25 asks you to build privacy in, and the cleanest way not to leak data to a processor is not to have one. No AI vendor on your processor list, no transfer assessments, no new data-processing agreement to negotiate.
On the box, inside your network, encrypted at rest. Every action your Aivell takes is written to the audit trail, and nothing is mirrored to anyone's cloud. Backups go wherever you decide they go.
An on-premise AI colleague reads every supplier invoice — PDF, scan or photo — and books it into your sheet or database. Invoices never leave your office.
An on-premise AI colleague reads each incoming contract, logs parties, value and renewal terms in a register — and books the deadlines on your calendar in time to act.
Applications screened against your job description into a ranked shortlist — read entirely on a box in your office, so candidate data never reaches a cloud.
Leave, benefits, expenses — employees ask in chat and get answers quoting your own HR handbook. Questions never leave the box in your office.
Every manual, contract and memo indexed on a box in your office. Your team asks in plain language and gets cited answers — nothing leaves the building.
Type a question in the chat, get the answer from your own SQL data — queried read-only, computed onboard, with the query shown for the curious.
An Aivell is an on-premise AI colleague: a small box that plugs into your network, runs its own AI onboard and takes the repetitive work off your desk. Prompts, documents, data — nothing ever leaves your office. Unplug the internet: it keeps working.
You describe each task as a chore, in plain language. Your Aivell turns it into a solid, guarded procedure and runs it in the background — with guardrails on every tool, a complete audit trail, and approvals in your hand when you want them. One-time setup, fixed monthly fee. No tokens, no overages.
onboard inference · air-gapped · fixed monthly fee