Building AI Agents on Red Hat AI

A hands-on tutorial for building, deploying, and extending AI agents on Red Hat AI (OpenShift AI on OpenShift) using the fips-agents toolkit.

Start the tutorial

What you'll build

A complete AI agent system deployed on Red Hat AI:

• An agent that solves calculus problems using remote math tools
• An MCP server providing calculus tools (integration, differentiation, limits, etc.)
• A gateway that proxies OpenAI-compatible requests to the agent
• A chat UI for interacting with the agent in a browser

Prerequisites

The full checklist (cluster, OpenShift AI, LLM serving, CLI tools, registry) lives in the Before You Begin module.

The tutorial supports two paths:

• Path A — OpenShift cluster with OpenShift AI and a GPU serving RedHatAI/gpt-oss-20b on vLLM (the full experience).
• Path B — no cluster GPU? Use the Developer Sandbox or any cluster, and point the agent at an external OpenAI-compatible model URL.

Example projects

Directory	Description
calculus-agent/	Complete agent configured for MCP-based calculus tools
calculus-helper/	FastMCP server with 8 SymPy-powered calculus tools

Local development

pip install mkdocs-material
mkdocs serve

Then open http://localhost:8000 to preview the tutorial.