Foundry Infrastructure — Deployments, Connections & Indexes

You can query the Foundry project infrastructure to discover available model deployments, data connections, knowledge indexes, and datasets.

Endpoint

All requests: http://localhost:8443 with ?api-version=2025-11-15-preview. Auth is automatic.

Operations

List model deployments

curl -s 'http://localhost:8443/deployments?api-version=2025-11-15-preview'

Returns all deployed models with name, publisher, version, SKU, and capabilities.

List connections

curl -s 'http://localhost:8443/connections?api-version=2025-11-15-preview'

Returns project connections (Azure AI Search, Bing, storage, etc.) with type and target URL.

List knowledge indexes

curl -s 'http://localhost:8443/indexes?api-version=2025-11-15-preview'

Returns available search indexes (Azure AI Search, Cosmos DB) for RAG scenarios.

List datasets

curl -s 'http://localhost:8443/datasets?api-version=2025-11-15-preview'

Returns datasets used for evaluation, fine-tuning, or agent training.

Get insights

curl -s 'http://localhost:8443/insights?api-version=2025-11-15-preview'

Returns evaluation insights and cluster analysis results.

When to use

• Discovering which models are available: "what models can I use?"
• Checking infrastructure: "what connections are configured?"
• Finding knowledge bases: "what indexes exist for search?"
• Listing datasets for evaluation or fine-tuning

When NOT to use

• For running inference (use /v1/chat/completions)
• For creating resources (that requires Azure Portal or Bicep)