Getting Started with Redis Iris

A hands-on tutorial for Redis Iris, the context and memory platform that gives AI agents real-time, compounding, retrieval, and navigable access to enterprise data.

This repo walks you through three of the core Iris tools using a sample healthcare dataset and a set of Jupyter notebooks:

Tool	What you'll do
LangCache	Semantically cache and reuse LLM responses to cut cost and latency.
Agent Memory	Store short-term session state and long-term durable memory.
Context Retriever	Model business entities and auto-generate MCP tools agents can call.

This README is the quick reference for setting up with Python and installing Jupyter.

NOTE: If you'd rather not install anything on your local machine, I recommend following the full Getting started with Redis Iris tutorial for detailed walkthrough of this repo using Google Colab.

Prerequisites

• Python >= 3.11
• Jupyter
• A Redis Cloud account
• Redis Insight (optional, for inspecting data)

Setup

1. Clone the repo

   git clone https://github.com/redis-developer/getting-started-with-redis-iris.git
   cd getting-started-with-redis-iris

2. Create your environment file

   cp .env.example .env

   Fill in the values as you create each Iris service in the Redis Cloud console:

   LANGCACHE_ENDPOINT=
   LANGCACHE_ID=
   LANGCACHE_KEY=
   
   AGENT_MEMORY_ENDPOINT=
   AGENT_MEMORY_STORE_ID=
   AGENT_MEMORY_KEY=
   
   CONTEXT_RETRIEVER_AGENT_KEY=

3. Set up a Python environment and install Jupyter

   In your terminal, create a virtual environment and install Jupyter.

   python -m venv .venv
   source .venv/bin/activate
   pip install --upgrade pip
   pip install jupyter

4. Create a Redis Cloud database and load data

   In the Redis Cloud console, create a new database. Hint: Follow these Setup instructions for more details.

   Then use Redis Insight to load the sample healthcare dataset data.redis. Hint: Follow the Context Retriever section for more details.

Notebooks

Each notebook installs its own dependencies (langcache, redis-agent-memory, redis-context-retriever, python-dotenv) in its first cell.

Launch Jupyter and open a notebook from the jupyter/ directory:

jupyter notebook

Notebook	Covers
jupyter/langcache.ipynb	Saving and semantically searching cached LLM responses.
jupyter/agent-memory.ipynb	Adding session events and long-term memories, then retrieving them.
jupyter/context-retriever.ipynb	Listing and calling the MCP tools generated from your entities.

Each service we cover in this repo also has its own setup and cleanup steps in the Redis Cloud console that's not in the notebooks. Hint: For each tool, follow the corresponding section in the full tutorial.

Repository layout

.
├── cli-instructions.md  # Context Retriever CLI (ctxctl) reference
├── data.redis           # Sample healthcare dataset (load via Redis Insight)
├── jupyter/             # The three walkthrough notebooks
└── .env.example         # Template for service credentials

Sample dataset for Redis Context Retriever

This tutorial uses a hospital management dataset with five entities — Patient, Doctor, Appointment, Treatment, and Bill — which map to Redis keys like treatment:{id} and appointment:{id}.

Context Retriever turns these into MCP tools such as get_doctor_by_id, filter_appointment_by_patient_id, search_treatment_by_text, and find_treatment_by_cost_range.

References

• Redis Iris Docs
• Blog: Context Is All You Need
• Getting started with Redis Iris