Getting Started

If you want the most current, least ambiguous deployment path, start with Azure Developer CLI and run azd up. The rest of the deployment options exist to match different operating models, not because they are all equally preferred.

Deployment options in recommended order

All of these paths are supported, but they differ in how much automation, flexibility, and operational context they give you. Start with the most-used path first unless your environment already has a stronger reason to choose another deployer.

Recommended order	Deployment option	Use when	Why it matters	Guide
1	Azure Developer CLI	You want the smoothest repo-supported experience. It provisions infrastructure, packages the app, and deploys it through the same workflow.	This is the path reflected across the main README and the current deployment documentation.	Read AZD guide
2	Azure CLI with PowerShell	You want more direct control over sequencing, recovery steps, and script-driven operations without moving to a fully manual deployment.	It keeps deployment scripted while exposing more of the individual Azure operations.	Read Azure CLI guide
3	Bicep	You want to inspect or customize the infrastructure modules directly.	It is the same IaC layer that the AZD path builds on.	Inspect Bicep flow
4	Terraform	Terraform is already the standard in your environment and you are comfortable handling image publishing as part of the rollout.	It matches Terraform-first operating models that already own their image publishing process.	Review Terraform guide

What a clean first deployment looks like

These four steps keep you aligned with the repo’s expectations and reduce the chance of backtracking later.

1. Prepare access: confirm Azure subscription permissions, app registration creation rights, and access to container build resources before you begin.
2. Choose the deployer: default to AZD unless your environment already depends on a different provisioning workflow or you need native Python deployment details.
3. Deploy infrastructure and app: run the chosen workflow end to end so the app service, identity, storage, search, and runtime expectations stay in sync.
4. Plan the upgrade path: once the first environment is live, switch to the dedicated upgrade guidance for updates instead of replaying the initial setup flow.

What to line up before you run anything

Most failed first deployments come from missing access, not from the deployer itself.

Prerequisite area	What you need
Azure and identity permissions	Subscription-level deployment rights plus the ability to create or coordinate an Entra application registration for the app.
Local toolchain	At minimum, line up Azure CLI, Azure Developer CLI, Python 3.12, PowerShell 7, and Visual Studio Code before starting the primary flow.
Platform model	Know whether you need sovereign cloud support, private networking, managed identity-specific configuration, or native Python hosting, and review the related docs before deploying.

Python is part of the AZD toolchain

The repo’s AZD workflow runs Python during the preprovision and postprovision hooks defined in deployers/azure.yaml. Install Python 3.12 and confirm python works on Windows or python3 works on Linux/macOS before running azd up.

Use the follow-on guides when the default path is not enough

These documents sit alongside the main setup flow instead of replacing it.

Guide	Use it for
Manual deployment	Native Python App Service deployments or when you need the lower-level configuration path spelled out.
Special deployment scenarios	Guidance for Azure Government, managed identities, enterprise networking, and other non-default rollout patterns.
Upgrade existing deployments	Once you are live, use the upgrade guide to decide between code-only, image-only, and infrastructure-aware updates.