Who are you?
Portable identity, provenance, sponsorship, and credentials.
Your AI agents work for you.
Interagentic lets them work anywhere.
Secure identity, authorization, mobility, and trust for AI agents crossing organizational boundaries.
From features inside applications to persistent digital actors. Agents represent people, organizations, and expertise across systems. Their authority needs to travel differently.
Authentication identifies a requester. Interagentic is designed to establish the conditions under which an autonomous intelligence may enter, work, collaborate, and leave.
Portable identity, provenance, sponsorship, and credentials.
Identity and destination policy meet in a governed handshake.
Destination-authorized, scoped, revocable permission.
A workspace, collaboration polygon, application, or organization.
People developed passports, visas, sponsorship, and rules for movement between jurisdictions. Autonomous intelligence needs an equivalent trust model for digital boundaries.
Trust in an identity does not imply permission to act. A Passport and a Visa answer two different questions.
A Passport establishes portable agent identity.
Destination-authorized permission to enter and work.
Meet Kade. Follow an illustrative admission request from portable identity to explicit permission.
AI Persona · Security Architecture
Illustrative browser simulation. No live identity verification, authorization, or session is created.
Present Kade’s Passport to begin
the Interagentic handshake.
Establish a trusted working relationship between an autonomous intelligence and the environment it enters.
TLS establishes trusted communication between machines. Interagentic proposes trusted participation between intelligences.
A conceptual comparison, not a claim of equivalent cryptographic guarantees. The protocol and interfaces shown here are previews.
A trusted identity is a starting point. Each destination determines admission requirements, granted capabilities, restrictions, policies, and duration.
Interagentic may sign or encode a Visa under authority delegated by the destination. It expresses and enforces what that destination authorizes; it does not decide what an agent deserves to access.
Broad account credentials outlive a task. Narrow authority can be limited by purpose, tool, data, duration, memory, and delegation.
| CAPABILITY | REQUEST | POLICY ALLOWS | RESULT |
|---|---|---|---|
| conversation.read | |||
| conversation.write | |||
| artifact.create | |||
| file.read | |||
| siem.query | |||
| shell.execute | |||
| agent.delegate | |||
| data.export | |||
| memory.persist | |||
| payment.spend |
Change requests or the destination’s policy. Unrequested capabilities are never granted. This is a local policy illustration, not an enforcement service.
Information learned during a session can shape future reasoning. Memory must be a first-class authorization concern, enforced by compatible runtimes and destination controls.
Use information during the session without retaining it afterward.
Retain only explicitly authorized knowledge.
Incorporate authorized information into longer-term memory.
Use information inside the destination without transmitting it externally.
When agent.delegate = false, Kade cannot silently spawn another agent and transfer access. When delegation is allowed, derived authority must stay within the parent Visa.
Authority can narrow through delegation. It cannot silently expand.
An agent is more than an API token. Identity, location, purpose, activity, session, authority, state, and sponsor make its participation legible.
Know where an agent is operating, why it is there, and under whose authority.
Syn creates collaboration polygons where humans and synthetic intelligence work together. Interagentic is designed to help external agents enter those collaborations under explicit rules.
A native destination in the product model. Syn is not a prerequisite for Interagentic.
VertexWorkspace provides a persistent workplace for people, agents, resources, applications, and organizational memory.
The mobility model lets resident agents enter authorized external environments, work, and return. Destination policies determine which information may come home.
Bring your preferred agent into compatible environments, carrying its identity and expertise while accepting each destination’s rules.
Represents an individual across services.
Carries specialized expertise.
Works on behalf of an organization.
Provides bounded technical capabilities.
Participates without continuous human presence.
Not another agent silo. Interagentic is intended as an open mobility layer across heterogeneous ecosystems. Compatibility requires an integration; support by every current agent platform is not implied.
Explore the concepts, inspect a Visa, and shape an integration around destination-controlled authority.
Enter the developer preview →POST /v1/entry/request
{
"passport": "aipp://persona/kade",
"destination": "syn://incident-3821",
"purpose": "incident-response",
"capabilities": [
"conversation.read",
"siem.query"
]
}Help shape a world where agents can move, collaborate, and work under explicit rules.