Overview

Datafor AI Agent allows administrators to assign usage permissions for specific Large Language Models (LLMs) to selected users and roles. Through the Permission Management interface, you can:

• Control which users or roles are allowed to use a given LLM
• Assign different permission levels (Read / Edit / Delete / Full Control)
• Manage or revoke access at any time

This guide explains how to view, add, modify, and remove permission assignments for an LLM.

1. How to Access the LLM Permission Management Interface

1. Open the left navigation menu and go to AI Agent → LLM
2. Select an LLM from the model list (e.g., qwen3-max)
3. Click the More Actions (…) icon beside the model
4. Choose Permission Management

The permission management window for the selected LLM will then open.

2. Understanding the Permission Management Interface

After entering the permission management window, you will see all users and roles currently authorized to access the selected LLM.

Top Action Bar

• Search Bar – Search by username, role name, email, etc.
• Export / Import – Export or import permission configurations (if enabled)
• + Add – Add new users or roles to the permission list

List Fields

Field	Description
Authorized Object	Name of the user or role granted access
Type	Indicates whether the object is a User or Role
Permission	The permission level assigned
Action	Button to remove the authorization

3. How to Add a New Authorized User or Role

Click the + Add button to open the add-permission dialog.

1. Use the search bar to locate a user or role
   • Supports searching by username, role name, phone number, or email
2. Select the permission level you want to assign:

Permission Level	Description
Read	Allows using the LLM (within assigned quota)
Edit	Allows editing LLM settings
Delete	Allows deleting the LLM
Full Control	Grants full administrative rights, including managing permissions

1. Select one or multiple users/roles
2. Click Confirm to complete the assignment

4. Modifying or Removing Permissions

Removing Permissions

To revoke access, click the × icon at the end of the corresponding row in the permission list. The user or role will immediately lose access to the LLM.

Modifying Permission Levels

The current version uses a “remove then reassign” approach for permission changes:

1. Remove the existing permission entry
2. Click + Add again and assign the new permission level

5. Recommendations for Assigning Permissions to Users vs. Roles

To manage permissions efficiently, especially in larger organizations, consider the following best practices:

Scenario	Recommended Approach
Large organizations	Prefer role-based authorization to avoid configuring many individual users
Sensitive or high-value models	Grant Full Control only to administrators
Regular business users	Assign Read to prevent accidental configuration changes
Report builders or modeling teams	Use roles such as Power User or Business Analyst for consistent access control

Note: LLM permissions do not inherit from report or folder permissions. They must be configured independently.