Liferay AI Hub

Search Bar

2026.08.19 - AI Hub Release

AI Hub Generally Available

We are excited to announce General Availability of Liferay AI Hub.

Liferay AI Hub is Now Generally Available
Liferay AI Hub is now generally available, giving organizations a low-code way to build, configure, and manage AI agents with speed and simplicity. Leverage Liferay AI Hub’s pre-built agents or create custom agents tailored to the unique requirements of your organization.

  • Deploy with confidence: Agents managed through Liferay AI Hub inherit your organization’s existing governance model. Agents can only perform tasks permitted to their users.

  • Test for free and scale as needed: Liferay AI Hub offers a 30-day free trial and flexible pricing tiers that can be configured to meet the specific needs of your organization.

 Content Gap Analyst

 

 AI Learn Assist Configuration
 

 AI Auto Tagging

Liferay AI Hub is now available (June 10th), the fastest way to bring AI-powered experiences into any digital environment.

AI Hub is a standalone, low-code SaaS product for building, configuring, and managing AI agents. Teams get a visual builder to create custom agents from scratch, plus a library of ready-to-use, Liferay-managed agents, like a writing assistant, for common business needs. No DXP subscription required: customers can deploy chatbots on any website, connect agents to external systems, and start using Liferay-managed agents independently of the rest of the Liferay stack.

AI Hub launched with two tiers, Activate and Enterprise, along with a set of capabilities that give administrators more control and agents more context to work with:

  • AI Guardrails: Powered by an integration with Google Model Armor, guardrails inspect both user input and AI output to protect against prompt injection attacks and prevent the agent from leaking personal or sensitive information.

  • AI Instructions: Custom instructions let administrators shape how an agent behaves and responds, tailoring tone and behavior to specific business needs.

  • Chat agent as a Click-to-Chat provider: Customers can power Click-to-Chat with AI agents, giving visitors instant, conversational answers instead of waiting on a live agent.

  • External website context for RAG: Agents can pull context directly from external websites, improving answer accuracy and relevance without manual content upload.

Release Notes 

 

AI Hub User Role Management

AI Hub account administrators can now manage their own team directly from the AI Hub interface, without involving Liferay provisioning or accessing the DXP Control Panel. Administrators can invite users by email, and invited people are automatically associated with the correct account when they accept. A set of predefined account roles governs who can manage agents, run them, configure data sources, and view activity, so organizations can structure access from the moment they onboard.

Key Benefits: 

  • Lets customer teams onboard and organize their own users without IT or vendor involvement

  • Speeds up team setup with email invitations and automatic account association

  • Gives administrators clear control over who can do what inside their AI Hub account

  • Reuses existing account-role infrastructure, lowering provisioning overhead

 

Advanced AI Guardrails

Gives administrators fine-grained control over AI agent behavior with configurable guardrails that keep agents from responding outside approved topics or producing inappropriate content.

Key Benefits: 

  • Reduces risk of AI agents giving off-brand or non-compliant responses

  • Gives administrators policy-level control without custom code

  • Builds end-user trust in AI-powered interactions

 

Liferay default AI Instructions

Ships AI agents with sensible default instructions out of the box, so teams get a working, well-behaved agent immediately instead of writing a prompt from scratch.

Key Benefits: 

  • Speeds up time-to-value for new AI agents

  • Reduces the expertise needed to configure AI behavior

  • Provides a consistent baseline experience across the platform

 

Agent Builder Permissions Management

Administrators can now control who can view, edit, and publish AI Agents through dedicated permissions in Agent Builder, keeping AI capabilities in the right hands.

Key Benefits: 

  • Enforces governance over who can create or modify AI agents

  • Reduces risk of unauthorized changes to agent behavior

  • Supports enterprise access-control requirements

 

Support a chat agent as Click-to-Chat provider

  • Feature Status: General Availability (GA)
  • Ticket: LPD-76586
  • Official Documentation: Chatbots

Customers can now power their Click-to-Chat experience with an AI chat agent, giving visitors instant, conversational answers instead of waiting for a live agent.

Key Benefits: 

  • Improves response times for customer inquiries

  • Reduces load on live support and sales teams

  • Provides a modern, conversational engagement channel

 

Provide context from external websites for improved RAG experiences

AI agents can now pull context directly from external websites, improving the accuracy and relevance of their answers without manual content uploads.

Key Benefits: 

  • Improves answer accuracy for AI agents

  • Reduces manual work to keep agent knowledge current

  • Expands the content sources agents can draw from

 

AI Instructions

Administrators can define custom instructions that shape how an AI agent behaves and responds, tailoring it to specific business needs and tone.

Key Benefits: 

  • Lets teams tailor AI behavior without code

  • Improves consistency of agent responses with brand voice

  • Provides a foundation for more advanced agent customization

 

AI Task Usage Metrics and Cost Tracking in Liferay Tokens

Administrators get visibility into how much AI usage is costing in Liferay Tokens, with usage metrics broken down by task, to help monitor spend and plan budgets.

Key Benefits: 

  • Gives finance and IT visibility into AI consumption costs

  • Helps avoid unexpected usage overages

  • Supports informed budgeting decisions for AI initiatives

 

Agents Monitoring and Auditing

 

Administrators can now monitor AI agent activity and review an audit trail of agent actions, supporting oversight and troubleshooting.

Key Benefits: 

  • Improves operational visibility into AI agent behavior

  • Supports compliance and audit requirements

  • Speeds up troubleshooting of unexpected agent responses

 

Allow customizing the chatbot Avatar

 

  • Feature Status: General Availability (GA)
  • Ticket: LPD-85257
  • Official Documentation: Chatbots

Businesses will be able to personalize their AI chat experience by customizing the chatbot's avatar to match their brand identity.

Key Benefits: 

  • Reinforces brand identity in AI-powered conversations

  • Improves visual consistency across customer touchpoints

  • Simple, no-code customization for business users

 

Create provisioning to connect AI Hub with DXP

 

Establishes the provisioning layer that connects a customer's DXP environment to AI Hub, enabling the rest of the AI Hub feature set to function.

Key Benefits: 

  • Enables DXP customers to connect to AI Hub

  • Provides the technical foundation for all other AI Hub capabilities

  • Reduces setup friction for new AI Hub customers

 

Feature Status Changes


AI Hub Homepage for logged users

 

  • Feature Status: Beta >>> General Availability (GA)
  • Ticket: LPD-90054
  • Official Documentation: AI Hub

The AI Hub homepage has moved from Beta to General Availability. Logged-in users now get a stable, fully supported homepage experience when they access AI Hub. No functional changes are expected for existing Beta users.

 

Make web crawler scalable

  • Feature Status: Beta >>> General Availability (GA)
  • Ticket: LPD-86162

The AI Hub web crawler used to build agent knowledge bases has moved from Beta to General Availability, with performance and scalability improvements to support larger content sources.

 

Establish a mechanism to have hard limits on Vertex consumption

Administrators can now set hard limits on AI model consumption, preventing runaway usage and unexpected costs. This capability has moved from Beta to General Availability.

On this page