LLMs are powerful, but they hallucinate and the real superpower comes when you give your AI agents tools and real-world context. In this video, I show you exactly how to do that with Embabel, a framework for building agentic flows on the JVM.

Building on our blog writing agent from the previous video, we expand the agent with new actions like TLDR generation, front matter creation, and web research. Along the way, you'll learn how to define custom tools (like a reading time calculator), wire in MCP servers (using Brave Web Search), and leverage Embabel's built-in tool groups to give your agents real capabilities beyond what the LLM knows on its own.

- How to define and register custom tools in Embabel using @LLMTool annotations
- How to configure and use MCP servers (e.g., Brave Web Search) for real-time web research
- How Embabel's built-in tool groups (web search, GitHub, etc.) work and how to apply them
- How goal-oriented action planning works — Embabel figures out the execution order from types
- Fun logging personalities (Star Wars, Severance) that add flair to your agent's console output

If you learned something new, give this video a thumbs up, subscribe to the channel, and drop a comment telling me what actions you'd add to your own blog writing agent!

0:00 - Intro & What We're Building Today
1:15 - Recap: Blog Agent from Part 1
2:30 - GitHub Repo & Starter Code
3:30 - Logging Personalities (Star Wars, Severance)
4:30 - Adding TLDR Summary Action
6:00 - Creating a Custom Reading Stats Tool
8:30 - Adding Front Matter with Tool Integration
10:30 - Setting Up Brave Search MCP Server
12:00 - Research Topic Action with Web Search Tool Group
14:00 - Running the Full Agent End to End
16:00 - Reviewing the Generated Blog Post & Front Matter
17:30 - Wrap Up & What's Next

🔗Resources & Links mentioned in this video:
📦 Blog Agent GitHub Repository: https://github.com/danvega/blog-agent
📖 Embabel Documentation - Tools: https://docs.embabel.com

👋🏻Connect with me:
Website: https://www.danvega.dev
Twitter: https://twitter.com/therealdanvega
Github: https://github.com/danvega
LinkedIn: https://www.linkedin.com/in/danvega
Newsletter: https://www.danvega.dev/newsletter

SUBSCRIBE TO MY CHANNEL: http://bit.ly/2re4GH0 ❤️

Dan Vega

A hands-on walkthrough of building a Java AI blog-writing agent using the Embabel framework on the JVM. Covers adding custom tools (a reading stats calculator), integrating MCP servers (Brave web search), using built-in tool groups for web research, and configuring logging personalities. The agent is extended step-by-step from two actions to a full pipeline that researches topics, writes drafts, adds TL;DR summaries, and generates front matter with reading time estimates.

Embabel Tools & MCP Servers: Supercharge Your Java AI Agents