Redis Iris Workshop

Build a smarter, faster AI agent with Redis Iris hands on, step by step.

Workshop Guide → redis-iris-workshop.vercel.app

What is this?

A hands on workshop where you pick an industry vertical and build an AI support agent powered by Redis Iris. You edit 5 exercise files. Everything else is pre-built.

Pick a vertical:

Vertical	Agent	Domain Key
Digital Native	Food delivery support	digital-native
Healthcare	Patient portal assistant	healthcare
Retail	Electronics shopping concierge	retail
Finance	Stock research analyst	finance
Banking	Customer care agent	banking
Telco	Wireless account support	telco

Quick Start

macOS / Linux

git clone https://github.com/Redislabs-Solution-Architects/redis-iris-workshop.git
cd redis-iris-workshop
make install
cp .env.example .env   # fill in Redis + OpenAI credentials
make seed-data
make dev

Windows (PowerShell)

git clone https://github.com/Redislabs-Solution-Architects/redis-iris-workshop.git
cd redis-iris-workshop
.\workshop.ps1 install
Copy-Item .env.example .env   # fill in Redis + OpenAI credentials
.\workshop.ps1 seed-data
.\workshop.ps1 dev

Note: If you get a "running scripts is disabled" error, run Set-ExecutionPolicy RemoteSigned -Scope CurrentUser once and try again.

Open localhost:3040 and follow the workshop guide.

What You Build

Each module adds one Redis Iris capability to your agent:

#	Module	What You Add	Exercise File
0	Setup	Redis Cloud database + environment	—
1	Vector Search	Search policy documents by meaning	exercises/<domain>/vector_search.py
2	Semantic Router	Block off topic queries before the LLM	exercises/<domain>/semantic_router.py
3	Context Retriever	Query live business data	exercises/<domain>/context_retriever.py
4	LangCache	Cache responses so repeats return instantly	exercises/<domain>/langcache.py
5	Agent Memory	Remember users across conversations	exercises/<domain>/agent_memory.py

Project Structure

exercises/<domain>/             ← You edit these (5 files per domain)
exercises/<domain>/solutions/   ← Reference implementations
backend/app/bases/              ← Pre-built boilerplate
backend/app/services/           ← Dynamic import layer (don't edit)
backend/app/                    ← FastAPI + LangGraph agent
frontend/                       ← React + Vite
domains/<domain>/               ← Schema, prompts, branding per vertical
scripts/                        ← Data seeding

Commands

On Windows, replace make <target> with .\workshop.ps1 <target>.

Command	Description
make install	Install all dependencies
make dev	Run backend + frontend
make seed-data	Module 0 — Load policies into Redis for Simple RAG
make setup-surface	Module 3 — Create Context Surface + agent key
make load-data	Module 3 — Load all entities via Context Surfaces
make seed-langcache	Module 4 — Seed one LangCache entry
make seed-memories	Module 5 — Seed long term memories
make status	Check which modules are active (server must be running)
make flush-redis	Flush all keys except memory
make reset	Flush Redis + re-seed policies

Prerequisites

• Python 3.11+ and uv
• Node.js 18+ and npm
• Redis Cloud account (free tier)
• OpenAI API key (provided by instructor)

Solutions

Set USE_SOLUTIONS=1 in your .env to run with all exercises pre-filled.