Table of Contents

Management Portal

The Management Portal is the administrative interface for configuring and managing your FoundationaLLM deployment.

Overview

The Management Portal provides IT administrators, developers, and power users with a comprehensive set of tools to:

  • Create and manage AI agents - Configure agent behavior, workflows, tools, shielding, and the chat experience
  • Configure models and tools - Manage AI model definitions, API endpoints, the tool library, and MCP servers
  • Connect and index data - Integrate data sources, build data pipelines, and manage knowledge units and vector databases
  • Control access - Manage role-based access control (RBAC), policy assignments, and impersonation
  • Enforce quotas - Define rate, usage, and resource limits and analyze quota health
  • Monitor usage - Review analytics dashboards and platform telemetry
  • Customize the platform - Manage themes, portal configuration, feature flags, packages, and deployments

Target Audience

The Management Portal is designed for:

Role Primary Tasks
IT Administrators Instance configuration, security settings, quotas, deployment management
Developers Agent creation, tool and workflow configuration, API integration
Power Users Agent management, prompt engineering, data source configuration

Getting Started

Quick Start

  1. Tour of the Management Portal - Learn the interface
  2. Creating Your First Agent - Build your first AI agent

Prerequisites

  • Microsoft Entra ID account
  • Access to a deployed FoundationaLLM instance
  • A role assignment that grants Management Portal access (see Management Portal Access)

Key Features

Agent Management

Create and configure AI agents with:

  • Workflows backed by workflow resources and Python plugin packages
  • Tools from the tool library, including knowledge, code, image generation, web grounding, and MCP server tools
  • Main workflow prompts and reusable prompt resources
  • Shielding plugins that screen content before and after completions
  • Portal experience settings such as welcome messages and conversation starters

Learn more:

Data Management

Connect and process organizational data:

  • Configure data source connections
  • Build and run data processing pipelines
  • Manage knowledge units, knowledge sources, and vector databases
  • Explore indexed knowledge

Learn more:

Security and Access Control

Implement comprehensive security:

  • Role-based access control (RBAC) at instance and resource scope
  • Portal access for the Management Portal, Chat User Portal, and Agent Portal
  • Policy assignments on resource types
  • Impersonation settings for trusted service principals
  • Agent access tokens for API integration

Learn more:

Quotas and Analytics

  • Define rate limits, usage limits, and resource limits
  • Analyze quota health and events
  • Review platform, user, agent, and token usage analytics

Learn more:

Platform Configuration

Customize and operate your deployment:

  • Themes for the Agent Portal and User Portal
  • Status message, welcome message, self-service settings, featured agents, and feature flags
  • Telemetry for errors and completions
  • API endpoints, packages, and managed deployments

Learn more:

Portal Sections

Section Description
Agents Agents, models, prompts, tools, and MCP servers
Analytics Platform, user, agent, and token usage dashboards
Data Data pipelines, pipeline runs, data sources, knowledge units, knowledge sources, vector databases, and Knowledge Explorer
Security Access control (role assignments), policies, and impersonation
Quotas Quota definitions and quota analysis
Platform Themes, configuration, telemetry, API endpoints, dynamic compute, packages, managed deployments, and sandboxes

See the Tour of the Management Portal for the pages in each section.

Reference Documentation