Voice AI · Made in Italy · Public beta
Fidus listens, answers out loud, reads your documents and tells you where every answer comes from. It is built for people for whom a screen is a wall — and for small businesses that deserve the kind of AI only large companies have today.
250 free seats · open to everyone
Why Fidus exists
Poor eyesight, shaky hands, or never having learned to use a computer: for millions of people apps are a wall, and everyday things mean asking someone for help.
Fidus is designed to be used by voice alone — ask, listen, done. During the beta we will test it with associations that support elderly and visually impaired people, and those associations will keep free accounts for good.
Large companies have AI that knows their documents, their numbers and their people. Small businesses usually don’t.
Fidus brings the same to them: it learns your documents, answers your team by voice and shows each person only what their role allows.
What it does
Speak naturally and Fidus answers out loud with a natural Italian voice. No menus to learn.
It reads your documents and tells you which one each answer comes from, so you can always check.
Nothing important happens without your confirmation, and every action is recorded.
In a team, each person sees only what their role allows. The system enforces it, not the AI.
Personal accounts, two-step verification for administrators and usage limits for each person.
Connect your email, calendar and documents — if and when you choose.
Already at work
Fidus is already running for a company with 17 locations, set up with roles that go from the owner to the staff of each site. Each person gets only the figures, documents and contacts their role allows.
The company’s name is not disclosed.
Public beta
Joining is free and works the same for everyone, whether or not you donate. Each person can ask up to 300 questions a month.
Leave your email and we will write to you when your seat is ready. When the seats are taken, you join the waiting list.
250 seats
Support the project
Servers, AI models and voice cost about €4,000 a month. We are raising £20,000 (about €24,000) on GoFundMe to keep Fidus running for six months.
A donation is a donation: it does not buy an account or any perk. The beta is free for everyone. Every month we publish what was spent and on what.
The GoFundMe page opens in the coming days.
Questions
No, on purpose. The beta is free and open to everyone, donors or not. Donations keep the service running for all.
Fidus is built by Aetherneum on top of Claude, the model by Anthropic, with its own voice, documents, roles and safeguards.
No advertising, and we never sell your data or use it to train models. Fidus runs on servers in Europe; understanding and speech rely on specialised AI providers under their business terms.
Fidus will have paid plans for individuals and businesses. Associations that support elderly and visually impaired people keep free accounts for good.
Giulio Gagliano, an Italian developer, as part of Aetherneum. Donations are received by Diamonds Hub Ltd, his UK company (Companies House 16437820), which pays for the servers and providers.