(01) Personal agentIn development

One

The personal agent that works for you.

(01) Understand

One is a personal agent that belongs to you, not to a model provider. It learns how you live and work through memory you can see and correct, keeps your tasks going across your devices, acts only within the limits you set, and interrupts you only when it is worth it.

(02) The question

“What should be done for you now, who should do it, and is it worth interrupting you?”

(03) Principles

  1. 01

    Remembers you, on your terms

    Every memory shows where it came from: something you said, something One observed or something it inferred. You can correct it, stop it being used or forget it for good.

  2. 02

    Acts within your limits

    For each kind of action you choose: do it, ask me first, or leave it to me. Authority comes from Access, never from a conversation.

  3. 03

    Protects your attention

    A small daily interruption budget, a clear “why now?” on every interruption, and a daily summary that ends. One is never built to keep you scrolling.

  4. 04

    Independent of any model

    Your memory, tasks, personality and permissions live outside the model. You never have to pick one, and your agent stays the same when models change.

  5. 05

    Private and yours

    Your keys stay on your devices and data is processed so that suiss cannot read it. It is never used for advertising, and you can export everything.

  6. 06

    Works with other people's agents

    Your One can propose, counter and agree with someone else's One. Agreements become commitments in Work.

(04) Capabilities

Memory

  • Source on every memory: said, observed or inferred
  • Correct, don't use, or forget for good
  • Life areas and your own projects
  • A screen that shows how One knows you
  • Full structured export

Tasks

  • Plans with steps, checkpoints and a clear finish
  • Official APIs and app actions first
  • A cloud browser that identifies itself
  • Long tasks that survive restarts
  • Hand-off to you when only you can do it

Autonomy

  • Act, act if pre-approved, ask, or hand off
  • Rules in plain language with a preview
  • “If this, do that, up to this amount, until this date”
  • Undo only where an action is truly reversible
  • What are you doing? Why? Stop. Undo. Don't do this again.

Attention

  • Decided on your device
  • Four interruption levels
  • A daily interruption budget you set
  • Waits for your next natural break
  • Quiet hours and focus modes as hard limits

Everywhere you are

  • iPhone and iPad: widgets, Live Activities, Shortcuts
  • Android: quick settings and live updates
  • Mac menu bar, Windows and Linux tray
  • Watch for status and quick approvals
  • CarPlay and Android Auto by voice
  • Turkish and English

(05) Agent-native

One plans the work, Access decides whether it may happen, and the action runs through a controlled executor. The model never holds your credentials.

  • Sub-agents that report back
  • Capabilities named by meaning
  • Browser extension and desktop helper
  • Bridge to your own coding agents
  • Content from emails and web pages treated as data, not instructions
  • Agent-to-agent agreements

(06) Guarantees

Never on its own

Some actions always need you, whatever your settings say.

  • 01Change passwords, two-factor or recovery settings
  • 02Add a payment method
  • 03Widen its own permissions
  • 04Delete your account or wipe its memory
  • 05Send money to a new recipient for the first time
  • 06Share sensitive data with someone new
  • 07Sign out or authorize devices
  • 08Install software

(07) Standards and protocols

  • MCP
  • A2A
  • App Intents
  • Web Bot Auth
  • WCAG 2.2 AA

(08) At a glance

Models
Operated by suiss, chosen for you per request
On device
Attention decisions, voice detection and an offline core
Privacy
Keys on your devices, processing suiss cannot read
Hardware
No suiss device required
Languages
Turkish and English
If One ever closes
At least 90 days' notice, then 30 days to export

(09) Works with

(10) Questions

What is suiss One?

A personal agent that belongs to you. It remembers what matters with memory you control, does tasks for you within limits you set, and works on your phone, computer, watch and car.

Which AI model does One use?

You don't have to choose. suiss operates the models and picks one per request. Your memory and permissions live outside the model, so One stays the same when models change.

Can One spend money or change my accounts?

Only within limits you approve through Access, and some actions, like adding a payment method or changing your password, always need you.

Is One available?

One is in development.