Agentic RAG with IBM Granite

Retrieval Agent

The Retrieval Agent is an implementation of Agentic RAG (Retrieval Augmented Generation) that can answer queries using a combination of both local document and web retrieval. It serves as a both personal productivity tool as well as an example of building an agent using task planning, adaptive step-by-step execution, and tool calling with Granite 3.2 Model at the helm. The core agent code is wrapped inside of an Open WebUI Function so that interaction with the agent can be accomplished through an easy to use chat UI.

Image Research Agent

The Image Research Agent is an AI-powered agent capable of conducting in-depth research based on image analysis that leverages CrewAI as an agentic framework. This tutotial uses the Granite 3.2 Vision Model alongside the Granite 3.2 8B Instruct Model, which offers enhanced reasoning capabilities.

Granite Agentic RAG

Build agentic RAG systems with Granite models

Granite Agentic RAG tutorial

Step-by-step tutorial for building agentic RAG

Introduction

Granite Language Cookbook

Granite Vision Cookbook

Granite Speech Cookbook

Granite Guardian Cookbook

Granite Time Series Cookbook

Granite Code Cookbook

Agentic AI Cookbook

Model Evaluation Cookbook

Additional Cookbooks

Agentic RAG with IBM Granite

Retrieval Agent

Image Research Agent

Granite Agentic RAG

Granite Agentic RAG tutorial

Introduction

Granite Language Cookbook

Granite Vision Cookbook

Granite Speech Cookbook

Granite Guardian Cookbook

Granite Time Series Cookbook

Granite Code Cookbook

Agentic AI Cookbook

Model Evaluation Cookbook

Additional Cookbooks

​Retrieval Agent

​Image Research Agent

Granite Agentic RAG

Granite Agentic RAG tutorial

Retrieval Agent

Image Research Agent