AI Studio Product API¶
The AI Studio Product API exposes the same product layer used by the AI Studio UI. Calls go through the product gateway, retain workspace and permission semantics, and create state that remains visible in the corresponding UI. Do not send Product API requests directly to the lower-level Catalog or workflow services.
Use these APIs to automate AI Studio resources such as workspaces, Catalog objects, connectors, workflows, knowledge bases, agents, model providers, SQL workbooks, compute resources, and alerts. For OpenAI-compatible model inference, use the Model API instead.
Before you call the API¶
Obtain the Product API base URL and a user-scoped API key or bearer token from the administrator of the target deployment.
Identify the workspace. Most resource calls require
X-Workspace-ID.Read Conventions for authentication, error handling, pagination, streaming, and retry rules.
Use the OpenAPI document published by the deployment for request and response schemas. Product versions and enabled modules can differ between deployments.
The following request shows the common envelope without assuming a deployment-specific resource payload:
curl "$MOI_PRODUCT_ENDPOINT/newmoi/catalog/healthz" \
-H "X-API-Key: $MOI_PRODUCT_API_KEY" \
-H "X-Workspace-ID: $MOI_WORKSPACE_ID"
Never use a system or core-service key in application code. A Product API key is evaluated by the product authentication, workspace authorization, and audit layers.
API areas¶
Area |
Use it for |
|---|---|
Discover identity, select a workspace, manage members, and inspect audit logs |
|
Browse Catalog, transfer files, configure sources, and run import or export tasks |
|
Manage workflow definitions and executions, rerun nodes, and inspect artifacts |
|
Manage semantic models and sources, then query selected data through Explore |
|
Discover an agent card and exchange A2A messages or task updates |
|
Configure LLM, embedding, and parser backends used inside AI Studio |
|
Manage SQL workbooks, inspect metadata, execute queries, and read history |
|
Provision worker capacity, inspect metrics, and manage alert rules and history |
Product routes are not a substitute for SQL privileges, external-system permissions, or provider credentials. The caller must satisfy those downstream controls in addition to AI Studio workspace authorization.