Build AI Agents the JAX-RS Way with Jakarta Agentic AI

One standard, annotation-driven model for agents in Java. Watch one get built live, run it yourself the same day, and swap the model with a single property.

Webinar

Hosted By

Luqman Saeed
Senior Java Developer, technical writer, product marketer, and author.

Luís Neto
Software Engineer II, Azul

Event Details

Dec 3, 2026

7am PT | 3pm GMT | 8:30pm IST

Every Java AI framework has its own idea of what an agent is. Jakarta Agentic AI aims to do for agentic applications what JAX-RS did for REST APIs in Java: one standard, annotation-driven model in place of each framework’s own.

An agent is a CDI bean, declared with annotations such as @Agent, @Trigger, @Decision and @Action, and it calls a model through an injectable LargeLanguageModel. The first milestone of the specification, 1.0.0-M1, came out on 30 July 2026 as a draft release for early community feedback. Payara Server Community 7.2026.8 includes a preview implementation, so you can run what you watch the same day, for free. In this webinar, Luqman Saeed and Luis Neto build an agent live against a local Ollama model, at no API cost.
You’ll see them:

Add a decision step and a tool call
Compose a second agent through a CDI event
Move from Ollama to Anthropic (Claude Opus) by changing a single MicroProfile Config property, with no recompile and no change to the code
Walk through the manual install step most people miss, so you don’t lose an evening to it

The demo closes with an overview of where the specification stands today. Jakarta Agentic AI 1.0 is still under development and is a standalone specification that sits outside the Jakarta EE platform for now. In M1, agents run synchronously, with no MCP support yet.

Who should attend:

Java and Jakarta EE developers
Payara users and platform engineers

Register Now

Trusted by 38% of the Fortune 100