Azure Container Registry client library for Python - version 1.2.0

Azure Container Registry allows you to store and manage container images and artifacts in a private registry for all types of container deployments.

Use the client library for Azure Container Registry to:

  • List images or artifacts in a registry
  • Obtain metadata for images and artifacts, repositories and tags
  • Set read/write/delete properties on registry items
  • Delete images and artifacts, repositories and tags

Source code | Package (Pypi) | Package (Conda) | API reference documentation | REST API documentation | Product documentation


Azure SDK Python packages support for Python 2.7 has ended 01 January 2022. For more information and questions, please refer to Python 3.7 or later is required to use this package. For more details, please refer to Azure SDK for Python version support policy.

Getting started

Install the package

Install the Azure Container Registry client library for Python with pip:

pip install --pre azure-containerregistry


To create a new Container Registry, you can use the Azure Portal, Azure PowerShell, or the Azure CLI. Here's an example using the Azure CLI:

az acr create --name MyContainerRegistry --resource-group MyResourceGroup --location westus --sku Basic

Authenticate the client

The Azure Identity library provides easy Azure Active Directory support for authentication. The DefaultAzureCredential assumes the AZURE_CLIENT_ID, AZURE_TENANT_ID, and AZURE_CLIENT_SECRET environment variables are set, for more information refer to the Azure Identity environment variables section

# Create a ContainerRegistryClient that will authenticate through Active Directory
from azure.containerregistry import ContainerRegistryClient
from azure.identity import DefaultAzureCredential

endpoint = ""
audience = ""
client = ContainerRegistryClient(endpoint, DefaultAzureCredential(), audience=audience)

Key concepts

A registry stores Docker images and OCI Artifacts. An image or artifact consists of a manifest and layers. An image's manifest describes the layers that make up the image, and is uniquely identified by its digest. An image can also be "tagged" to give it a human-readable alias. An image or artifact can have zero or more tags associated with it, and each tag uniquely identifies the image. A collection of images that share the same name but have different tags, is referred to as a repository.

For more information please see Container Registry Concepts.


The following sections provide several code snippets covering some of the most common ACR Service tasks, including:

Please note that each sample assumes there is a CONTAINERREGISTRY_ENDPOINT environment variable set to a string containing the https:// prefix and the name of the login server, for example "". Anonymous access samples are getting endpoint value from environment variableCONTAINERREGISTRY_ANONREGISTRY_ENDPOINT.

List repositories

Iterate through the collection of repositories in the registry.

with ContainerRegistryClient(self.endpoint, self.credential) as client:
    # Iterate through all the repositories
    for repository_name in client.list_repository_names():

List tags with anonymous access

Iterate through the collection of tags in the repository with anonymous access.

with ContainerRegistryClient(endpoint) as anon_client:
    manifest = anon_client.get_manifest_properties("library/hello-world", "latest")
    print(f"Tags of {manifest.repository_name}: ")
    # Iterate through all the tags
    for tag in manifest.tags:

Set artifact properties

Set properties of an artifact.

with ContainerRegistryClient(self.endpoint, self.credential) as client:
    # Set permissions on image "library/hello-world:v1"

Delete images

Delete images older than the first three in the repository.

with ContainerRegistryClient(self.endpoint, self.credential) as client:
    for repository in client.list_repository_names():
        # Keep the three most recent images, delete everything else
        manifest_count = 0
        for manifest in client.list_manifest_properties(
            repository, order_by=ArtifactManifestOrder.LAST_UPDATED_ON_DESCENDING
            manifest_count += 1
            if manifest_count > 3:
                # Make sure will have the permission to delete the manifest later
                print(f"Deleting {repository}:{manifest.digest}")
                client.delete_manifest(repository, manifest.digest)

Upload images

To upload a full image, we need to upload individual layers and configuration. After that we can upload a manifest which describes an image or artifact and assign it a tag.

self.repository_name = "sample-oci-image"
layer = BytesIO(b"Sample layer")
config = BytesIO(json.dumps(
        "sample config": "content",
with ContainerRegistryClient(self.endpoint, self.credential) as client:
    # Upload a layer
    layer_digest, layer_size = client.upload_blob(self.repository_name, layer)
    print(f"Uploaded layer: digest - {layer_digest}, size - {layer_size}")
    # Upload a config
    config_digest, config_size = client.upload_blob(self.repository_name, config)
    print(f"Uploaded config: digest - {config_digest}, size - {config_size}")
    # Create an oci image with config and layer info
    oci_manifest = {
        "config": {
            "mediaType": "application/vnd.oci.image.config.v1+json",
            "digest": config_digest,
            "sizeInBytes": config_size,
        "schemaVersion": 2,
        "layers": [
                "mediaType": "application/vnd.oci.image.layer.v1.tar",
                "digest": layer_digest,
                "size": layer_size,
                "annotations": {
                    "": "artifact.txt",
    # Set the image with tag "latest"
    manifest_digest = client.set_manifest(self.repository_name, oci_manifest, tag="latest")
    print(f"Uploaded manifest: digest - {manifest_digest}")

Download images

To download a full image, we need to download its manifest and then download individual layers and configuration.

with ContainerRegistryClient(self.endpoint, self.credential) as client:
    # Get the image
    get_manifest_result = client.get_manifest(self.repository_name, "latest")
    received_manifest = get_manifest_result.manifest
    print(f"Got manifest:\n{received_manifest}")
    # Download and write out the layers
    for layer in received_manifest["layers"]:
        # Remove the "sha256:" prefix from digest
        layer_file_name = layer["digest"].split(":")[1]
            stream = client.download_blob(self.repository_name, layer["digest"])
            with open(layer_file_name, "wb") as layer_file:
                for chunk in stream:
        except DigestValidationError:
            print(f"Downloaded layer digest value did not match. Deleting file {layer_file_name}.")
        print(f"Got layer: {layer_file_name}")
    # Download and write out the config
    config_file_name = "config.json"
        stream = client.download_blob(self.repository_name, received_manifest["config"]["digest"])
        with open(config_file_name, "wb") as config_file:
            for chunk in stream:
    except DigestValidationError:
        print(f"Downloaded config digest value did not match. Deleting file {config_file_name}.")
    print(f"Got config: {config_file_name}")

Delete manifest

with ContainerRegistryClient(self.endpoint, self.credential) as client:
    get_manifest_result = client.get_manifest(self.repository_name, "latest")
    # Delete the image
    client.delete_manifest(self.repository_name, get_manifest_result.digest)

Delete blob

with ContainerRegistryClient(self.endpoint, self.credential) as client:
    get_manifest_result = client.get_manifest(self.repository_name, "latest")
    received_manifest = get_manifest_result.manifest
    # Delete the layers
    for layer in received_manifest["layers"]:
        client.delete_blob(self.repository_name, layer["digest"])
    # Delete the config
    client.delete_blob(self.repository_name, received_manifest["config"]["digest"])


For infomation about troubleshooting, refer to the troubleshooting guide.


ACR client library will raise exceptions defined in Azure Core.


This library uses the standard logging library for logging.

Basic information about HTTP sessions (URLs, headers, etc.) is logged at INFO level.

Detailed DEBUG level logging, including request/response bodies and unredacted headers, can be enabled on the client or per-operation with the logging_enable keyword argument.

See full SDK logging documentation with examples here.

Optional Configuration

Optional keyword arguments can be passed in at the client and per-operation level. The azure-core reference documentation describes available configurations for retries, logging, transport protocols, and more.

Next steps


This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit

This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact with any additional questions or comments.