Skip to content

00 Prerequisites

Goal

Make sure you can start the hands-on labs without hidden setup gaps.

Estimated time

10 to 15 minutes if the Azure resources already exist.

Official references

Checklist

  • This repository is cloned locally, or you have forked it and cloned your fork, so the example files are available on your machine.
  • Azure subscription is active.
  • Azure CLI is installed and az login works.
  • Python 3.10 or newer is installed.
  • The workshop Python packages are installed from requirements.txt (see below).
  • You have an existing Foundry project or permission to create one.
  • You have a role such as Azure AI User on the project.

Verify locally

python --version
az version
az account show

Install the Python packages

All labs share one dependency set, so you only install once. From the repository root, create a virtual environment and install requirements.txt:

python -m venv .venv
source .venv/bin/activate
pip install --upgrade pip
pip install -r requirements.txt

This installs everything the lessons use (the Foundry SDK, authentication, tracing, YAML, and the docs tooling), so you do not need to run pip install again before each lesson. On Windows, activate the environment with .venv\Scripts\activate instead.

Environment variables used in this workshop

Primary names:

  • AZURE_AI_PROJECT_ENDPOINT
  • AZURE_AI_MODEL_DEPLOYMENT_NAME
  • AZURE_AI_AGENT_NAME
  • KEEP_AGENT

Optional later-lab name:

  • AZURE_AI_WORKFLOW_NAME

Fallback names accepted by the example code:

  • PROJECT_ENDPOINT
  • MODEL_DEPLOYMENT_NAME
  • AGENT_NAME

KEEP_AGENT controls whether workshop examples that create agents clean them up at the end. The default is true, which keeps created agents available in Foundry for later inspection and reuse. Set KEEP_AGENT=false if you want disposable demo runs instead.

AZURE_AI_WORKFLOW_NAME is only needed for the native workflow lab if you save the workflow under a different name than the workshop default.

Success criteria

You are ready to continue when:

  • Azure CLI can show your account
  • the workshop packages installed without errors
  • you know your project endpoint
  • you know the name of a deployed model