canonical: https://jentic.com/apis/azure.com/azure-machine-learning-workspaces-management-client # Microsoft Azure Machine Learning Workspaces Management Client Jentic publishes the only available OpenAPI specification for Machine Learning Workspaces Management Client, keeping it validated and agent-ready. This API manages Microsoft.MachineLearning workspaces - the legacy 'classic' Azure Machine Learning Studio workspace resource, which contains experiments, datasets, and web service plans. It exposes workspace provisioning, key listing, and resync via Azure Resource Manager. It does not run experiments or manage Azure Machine Learning v2 workspaces. ## For AI agents Provision, list, and manage classic Azure ML Studio workspaces and their access keys through Azure Resource Manager. Use for workspace lifecycle, not for running experiments or modern Azure ML v2. ## Scope Does not run experiments, manage compute, or operate Azure ML v2 (Microsoft.MachineLearningServices) - use only for classic Microsoft.MachineLearning workspace lifecycle and key operations. ## Capabilities - Provision and delete classic Azure Machine Learning Studio workspaces - List Microsoft.MachineLearning workspaces across a subscription or resource group - Retrieve and resynchronise workspace storage account keys - Update workspace tags and properties through PATCH - Inspect the operations catalog exposed by the Microsoft.MachineLearning provider ## Use cases ### Classic ML Studio Workspace Provisioning Data science teams running on Azure ML Studio (classic) provision new workspaces tied to a storage account. Workspaces_CreateOrUpdate (PUT on /workspaces/{workspaceName}) accepts a SKU, location, and ownerEmail, and returns the workspace resource. The ARM call is synchronous for classic workspaces, so the resource is ready immediately for downstream Studio access. Example prompt: PUT a workspace named 'demos-ml' in westeurope with ownerEmail 'team@example.com' and a Standard SKU to /subscriptions/{subscriptionId}/resourceGroups/ml-prod/providers/Microsoft.MachineLearning/workspaces/demos-ml. ### Workspace Key Rotation and Resync When a storage account behind a classic ML workspace has its keys rotated, the workspace's cached credential must be refreshed. ResyncStorageKeys forces the workspace to pull the current key from the linked storage account, restoring access for experiments without requiring a workspace rebuild. Example prompt: Call POST /subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.MachineLearning/workspaces/{workspaceName}/resyncStorageKeys for each workspace bound to the rotated storage account. ### Subscription Inventory and Decommission Planning Platform owners migrating off classic Azure ML Studio inventory every Microsoft.MachineLearning workspace in the subscription, then decide which to decommission. Workspaces_List returns the full set with creation date and SKU, and Workspaces_Delete removes the resource cleanly. Combined with tagging conventions, this drives a controlled migration to Azure ML v2. Example prompt: List all Microsoft.MachineLearning workspaces in the subscription created before 2024, return name and resource group, and DELETE any tagged decom=true. ### AI Agent Workspace Operations Maintenance agents that own a portfolio of classic ML workspaces use Jentic to find the right ARM operation on demand. Searching for a workspace intent returns the matching schema; the agent loads it and executes against management.azure.com with a token retrieved from your Jentic One instance - no SDK install required. Example prompt: Use Jentic to search 'list azure ml studio workspaces', load the Workspaces_List operation, and execute against the target subscription, returning workspace name and SKU per resource group. ## Key endpoints | Method | Path | Description | | --- | --- | --- | | PUT | /subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.MachineLearning/workspaces/{workspaceName} | Create or update a classic ML workspace | | GET | /subscriptions/{subscriptionId}/providers/Microsoft.MachineLearning/workspaces | List ML workspaces in a subscription | | DELETE | /subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.MachineLearning/workspaces/{workspaceName} | Delete a classic ML workspace | | POST | /subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.MachineLearning/workspaces/{workspaceName}/listWorkspaceKeys | List workspace storage keys | | PATCH | /subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.MachineLearning/workspaces/{workspaceName} | Update workspace tags or properties | ## Key resources - **Workspaces** — Classic Microsoft.MachineLearning workspace lifecycle: create, list, get, update, delete, list keys, resync keys. - **Operation** — Operations catalog exposed by the Microsoft.MachineLearning resource provider. ## Why Jentic - **Setup:** Wiring the classic Microsoft.MachineLearning workspace API by hand means registering an Azure AD app, negotiating OAuth 2.0 tokens for management.azure.com, and threading subscription and resource-group scope through every workspace call. Through Jentic you install once, import Machine Learning Workspaces Management Client from the API Directory, store the Azure AD credential once, and your agent calls it. - **Permission scoping:** This API puts the workspace in the URL path (/resourceGroups/{resourceGroupName}/providers/Microsoft.MachineLearning/workspaces/{workspaceName}), so a rule can pin your agent to one workspace: it can list and read workspaces. You choose the operations it may call, so listing workspace keys or deleting a workspace is not included unless you add it. - **Credential handling:** Your Azure AD credential for the classic MachineLearning control plane is stored once, encrypted, by your own Jentic One instance and injected at execution time. It never enters the agent's prompt, logs, or context. - **Discovery method:** Agents search Jentic by intent such as 'list classic ML workspaces' or 'read an ML workspace', and Jentic returns the matching operation with its input schema so the agent calls the right endpoint without browsing the reference docs. ## Related APIs - **Azure Machine Learning Workspaces** — The modern Microsoft.MachineLearningServices workspace surface for current Azure ML projects. - **Machine Learning Compute Management Client** — Manages legacy operationalization clusters that pair with classic ML workspaces. - **Azure Machine Learning Datastore Management Client** — Manage datastores that host data for ML experiments. ## FAQ ### Why is there no official OpenAPI spec for Machine Learning Workspaces Management Client? Microsoft Azure does not publish an OpenAPI specification. Jentic generates and maintains this spec so that AI agents and developers can call Machine Learning Workspaces Management Client via structured tooling. It is validated against the live API and kept up to date. Get started with Jentic One, the self-hosted execution layer. ### What authentication does the Machine Learning Workspaces Management Client use? AAD OAuth 2.0 bearer tokens scoped to https://management.azure.com/. Through Jentic, AAD client credentials live in the encrypted vault and a short-lived bearer token is injected at call time. ### Does this API manage Azure ML v2 workspaces? No. This API manages the classic Microsoft.MachineLearning provider (Azure ML Studio classic). The current Azure Machine Learning service uses the Microsoft.MachineLearningServices provider, which exposes a different workspace resource with online endpoints, compute clusters, and pipelines. ### What are the rate limits for the Machine Learning Workspaces Management Client? Standard Azure Resource Manager throttling: typically 12,000 reads and 1,200 writes per hour per subscription per region. Workspace create and delete are typically synchronous for classic workspaces. ### How do I rotate workspace storage keys through Jentic? Run pip install jentic, then search Jentic for 'resync azure ml workspace storage keys'. Jentic returns the Workspaces_ResyncStorageKeys operation; load it, supply the workspace path, and execute the POST against /workspaces/{workspaceName}/resyncStorageKeys. ### Can I run experiments or train models with this API? No. Experiment, dataset, and web service operations live in the ML Studio data plane (or in Azure ML v2 for modern workloads). This client is the ARM control plane for the workspace resource only. ### Can I limit what my agent is allowed to do with the Machine Learning Workspaces Management Client? Yes. Because this API puts the workspace in the URL path (/resourceGroups/{resourceGroupName}/providers/Microsoft.MachineLearning/workspaces/{workspaceName}), a rule in your self-hosted Jentic One instance can pin your agent to a single workspace and let it only list and read workspaces. You decide which operations the agent may call, so higher-impact actions like listing workspace storage keys or deleting a workspace stay off the table unless you explicitly add them. Your Azure AD credential is stored by your own instance and injected at call time, so the agent runs only the operations you allow with credentials it never sees.