Permissions Overview

Karyam uses Role-Based Access Control (RBAC) to ensure users have access only to the resources and capabilities required for their responsibilities.

Permissions are designed to help organizations securely manage AI development, operations, and consumption while maintaining clear separation between administrators, developers, and business users.

Why Permissions Matter

As organizations scale their AI initiatives, not every user should have access to every resource.

Permissions help you:

Control access to platform resources
Protect sensitive configurations and data
Govern who can create, modify, or publish AI capabilities
Enable secure collaboration across teams
Reduce operational and security risks

Permission Model

Karyam permissions are applied across different levels of the platform.

Workspace
    ├── Teams
    ├── Users
    └── Resources
            ├── Agents
            ├── AI Flows
            ├── Knowledge Bases
            ├── Models
            ├── Skills
            ├── MCP Servers
            └── Service Catalogs

Access to these resources is determined by a user’s assigned role and team membership.

Core Concepts

The permission model is built around several key concepts:

Roles

Roles define what actions a user can perform within the platform.

Typical responsibilities include:

Platform administration
AI development
Business consumption of published AI services

Learn more in the Roles guide.

Resource Permissions

Each resource has its own set of permissions that determine who can view, create, edit, delete, execute, or publish it.

Examples include:

Agents
AI Flows
Knowledge Bases
Models
Skills
MCP Servers
Service Catalogs

Team Access

Teams simplify permission management by allowing administrators to assign access to groups of users instead of managing permissions individually.

This makes collaboration easier while ensuring consistent access across projects.

Publishing Permissions

Only authorized users can publish Agents and AI Flows to the Service Catalog.

Publishing makes approved AI capabilities available to business users while keeping development resources private.

Best Practices

To maintain a secure and well-governed environment, consider the following recommendations:

Assign users the minimum permissions required for their role.
Use Teams to manage access at scale.
Restrict publishing to trusted users.
Separate development and production workspaces.
Periodically review user access and published resources.
Require Human Confirmation for sensitive or irreversible workflows.