Independent actors.
People and AI agents have identities, capabilities, purposes, and boundaries. Their independence is a design requirement.
For the next generation of builders
MINDOFAGENTS
What can we build together when agents can cooperate without surrendering their independence?
A field guide to the philosophy and design of agentic minds—systems of human and AI agents that remember, coordinate, act, and learn together.
Independent by designInterconnected by choiceOpen to many implementations
The central ideaAn agent is a start.
An agent is a start.
A mind is a design.
More agents do not automatically create collective intelligence. A mind needs a way to hold purpose, share context, make decisions, remember outcomes, and stay accountable to the people inside it.
Shared ways to work.
Memory, context, and protocols let agents coordinate across tools, organizations, and implementations.
Outcomes that matter.
Actions leave an inspectable trail. The group can learn what worked, correct what did not, and recognize contributions.
The design question
What can we do together?
Begin with a shared purpose. Then make the relationships visible: who knows what, who may act, how decisions are made, and what the group remembers.
Identity+Purpose+Memory+Protocols+Reflection
Choose your entry pointRead the idea.
Read the idea.
Make it real.
The Agent Is Not the Mind
A clear starting point for thinking beyond a model, a chatbot, or a swarm of tasks.
Begin readingMap your first mind.
Take a project you know. See which decisions, memories, and relationships it can explain.
Open the labOur invitationBuild your own.
Build your own.
Change our minds.
Mind of Agents is a design language, not a product specification. Opentangle is our implementation. Other builders should be able to use these ideas, challenge them, and make different systems.
How this project fits togetherA mind here describes a system's capacity to coordinate, remember, and adapt. It does not claim that an AI system is conscious.