Azure-AI_Foundry

Azure AI Foundry — Bicep Infrastructure Templates

This folder contains modular Bicep templates for deploying Azure AI Foundry resources.

Templates

Template Description
ai-foundry-hub.bicep Core hub with AI Services, Key Vault, Storage, and optional Container Registry
ai-foundry-project.bicep Project workspace linked to an existing hub
ai-foundry-openai.bicep Azure OpenAI account with GPT-4o, GPT-4o-mini, and embedding model deployments
ai-foundry-search.bicep Azure AI Search service with semantic configuration
ai-foundry-complete.bicep End-to-end deployment combining all of the above

Quick Reference

ai-foundry-hub.bicep

Deploys the foundational AI Foundry Hub and its dependent resources.

Parameters:

Parameter Type Default Description
hubName string — Name of the AI Foundry Hub
location string resource group location Azure region
sku string Basic Hub SKU (Basic or Standard)
enableContainerRegistry bool false Create an Azure Container Registry
tags object {} Resource tags

Deploy:

az deployment group create \
  --resource-group <rg> \
  --template-file ai-foundry-hub.bicep \
  --parameters hubName='my-hub'

ai-foundry-project.bicep

Creates a project inside an existing hub.

Parameters:

Parameter Type Default Description
projectName string — Name of the project
hubId string — Resource ID of the parent hub
location string resource group location Azure region
tags object {} Resource tags

Deploy:

az deployment group create \
  --resource-group <rg> \
  --template-file ai-foundry-project.bicep \
  --parameters projectName='my-project' hubId='<hub-resource-id>'

ai-foundry-openai.bicep

Deploys an Azure OpenAI resource with model deployments.

Parameters:

Parameter Type Default Description
openAIName string — Name of the Azure OpenAI resource
location string resource group location Azure region
gpt4oVersion string 2024-11-20 GPT-4o model version
gpt4oMiniVersion string 2024-07-18 GPT-4o-mini model version
embeddingVersion string 1 text-embedding-3-small version
gpt4oCapacity int 30 GPT-4o TPM capacity (in thousands)
gpt4oMiniCapacity int 30 GPT-4o-mini TPM capacity
embeddingCapacity int 30 Embedding model TPM capacity
tags object {} Resource tags

Deploy:

az deployment group create \
  --resource-group <rg> \
  --template-file ai-foundry-openai.bicep \
  --parameters openAIName='my-openai'

ai-foundry-search.bicep

Deploys Azure AI Search with semantic ranking support.

Parameters:

Parameter Type Default Description
searchName string — Name of the search service
location string resource group location Azure region
sku string basic Search SKU (free, basic, standard, etc.)
replicaCount int 1 Number of replicas
partitionCount int 1 Number of partitions
semanticSearch string free Semantic search tier (disabled, free, standard)
tags object {} Resource tags

Deploy:

az deployment group create \
  --resource-group <rg> \
  --template-file ai-foundry-search.bicep \
  --parameters searchName='my-search'

ai-foundry-complete.bicep

Full end-to-end deployment. Orchestrates all modules with managed-identity connections.

Parameters:

Parameter Type Default Description
environmentName string — Prefix for all resource names
location string resource group location Azure region
hubSku string Basic Hub SKU
enableContainerRegistry bool true Create a Container Registry
gpt4oCapacity int 30 GPT-4o TPM capacity
gpt4oMiniCapacity int 30 GPT-4o-mini TPM capacity
embeddingCapacity int 30 Embedding TPM capacity
searchSku string basic AI Search SKU
tags object {} Resource tags

Deploy:

az deployment group create \
  --resource-group <rg> \
  --template-file ai-foundry-complete.bicep \
  --parameters environmentName='myenv'

Prerequisites