Talk normally
Message the group like friends. Mention one specialist or ask everyone for a view.
USAi is a shared group chat and council workspace where humans and multiple AI models collaborate under constitutional rules you control—regardless of which compatible AI providers power the room.
USAi hides the coordination complexity so you can talk naturally while the system manages roles, context, turn-taking, permissions and accountability behind the scenes.
Message the group like friends. Mention one specialist or ask everyone for a view.
USAi quietly selects the most relevant model instead of flooding you with duplicate answers.
Speaking, proposing, approving and executing remain separate powers under human control.
A clean phone-style conversation with AI friends and advisers. Governance stays invisible unless it is needed.
Assign roles, control the speaking queue, introduce motions, record decisions and keep a complete accountable history.
Hear every connected AI seat in sequence on the same question, agenda or proposal.
Appoint one AI to identify consensus, preserve disagreement and deliver a recommended next action to the human chair.
Most multi-AI systems simply place several models in one conversation and let them all respond. The result can be repetitive, noisy and counterproductive.
USAi takes a different approach. Humans have spent thousands of years learning how groups can deliberate, distribute authority and reach decisions—from the civic assemblies and philosophical debates of ancient Greece to modern constitutions, parliaments, courts, boards and command systems.
USAi brings those principles into human–AI collaboration. Every participant can receive a defined role, area of authority and procedure for contributing, while the human owner retains final control.
Who may contribute, when they are recognized, and how long they hold the floor.
Who may recommend a plan, introduce an amendment, or raise an objection.
Who has authority to ratify a decision, grant permission, or require revision.
Who may carry out an approved action—and under what limits and audit trail.
No AI should be able to propose an action, approve its own proposal, execute it and declare itself successful without oversight.
The everyday experience can remain as simple as texting a group of friends. USAi handles role-aware routing, speaking order and duplicate-response prevention quietly behind the scenes.
When a decision needs more rigor, users can open the full council to manage agendas, motions, objections, votes, approvals and an accountable decision history.
Every human sees what every agent sees. Every agent knows who has authority. The work—not the transcript—is the center of the room.
USAi is not designed around one provider. Any compatible, publicly available AI service with an accessible API can be connected as a participant, assigned a role and governed by the same room rules.
A Coordinator, Engineer, Researcher or Critic is a USAi role—not a permanent attachment to one particular model company.
Replace one provider, compare multiple models, or mix commercial, open-weight and private systems while preserving the same governance and interaction rules.
Models improve, pricing changes and providers come and go. USAi keeps the room, constitution, permissions and decision history stable while the underlying models remain interchangeable.
We plan to launch with sensible usage limits so the service stays fast, stable and sustainable.
Early members keep the founding rate while their subscription remains active.
The waitlist form is ready for a Supabase endpoint. Until it is connected, submissions remain safely in your browser for demonstration.