Hire your next employee. In minutes.

A teammate, not a chatbot. Your AI employee has a name, a face, a voice, and a memory that persists — they remember last week's decision and the way you like things written. They have their own workspace: their own files, their own history with you, and the role you gave them.

You talk to them the way you'd talk to a good new hire. They do the work and bring it back for you to look at.

Desk work, the real kind. Writing and drafting — email, briefs, reports, investor updates, first-pass contracts. Research — going out to the web, reading what's there, and coming back with a sourced answer instead of a confident guess. Documents and spreadsheets. Inbox and calendar — triage, scheduling, chasing the replies you never got. And watching — keeping an eye on something you asked about and telling you the moment it moves.

What they can't do is anything you haven't handed them. Every connection — every mailbox, every channel, every folder — is one you switch on yourself.

We staff an AI employee the way you'd staff a person: by the shift. A standard hire works eight hours a day. During the shift they're on — watching their channels, picking up whatever you hand them, running the tasks in their queue, and raising anything that needs a human decision. Outside the shift they're off the clock and quiet.

It's a deliberate choice. Nobody budgets a coworker in tokens, and usage-metered AI makes you flinch every time you ask for something — which is exactly the wrong instinct to have about a teammate. A shift is one predictable line you can decide on before you commit.

Yes. The shift is anchored to a clock you pick, so you can run an employee overnight and find the work done in the morning, or stagger two of them across the day so someone is always on. If one desk needs more than eight hours, you add a second shift to it rather than a second person.

However you already work. Slack, Telegram, Gmail, Outlook — or face to face in their workspace, where you can just talk. No prompt engineering, no special syntax, no learning a tool. If you can brief a person, you can brief them.

Three things, and they're the reason the product exists. Anything consequential lands in an approval inbox before it leaves your company — you approve it, send it back for changes, or kill it. Every action they take is written to a run history you can read afterward, in plain language. And there is one stop button that halts everything, immediately.

They do the work. You keep the authority.

No. There are no API keys to paste, no models to choose, no infrastructure to stand up. Pick a face, give them a role, connect the channels you want them in, and start talking. It's minutes, not a project — the tutorial above is the whole thing, start to finish.

Each employee runs in an isolated workspace of their own — their own database, their own container — not a shared pool with other customers. Your files, their memory, and the whole history of your work together belong to you, and your work is not used to train models.

Yes — that's the point. Each hire is their own person with their own role, memory, and shift, and you see the whole roster in one manager view: who's on, what they're working on, and what's waiting on you. Start with one desk and grow the team the way you'd grow any other.

We haven't published pricing yet. During the private beta we'd rather get a small number of people all the way to real, useful work than sell seats — so beta access is free while we're still hardening the platform, and we'll talk about what comes after with each tester directly.

When pricing does go up, it will be one flat line per employee, per shift, per month, with the AI usage included. No meter, no surprise bill at the end of a good month.

Send us an email. We're opening seats a few at a time so that every tester gets a real onboarding instead of a signup form — tell us what you'd put an employee to work on and we'll come back to you.