Skip to content
AI + machine learning srch

AI Search naming convention

The CAF abbreviation for AI Search is srch. Names can be 2 to 60 characters long and allow lowercase alphanumerics and hyphens. The name must be globally unique across all of Azure.

Naming rules

CAF abbreviation srch
Length 2 to 60 characters
Allowed characters Lowercase letters and numbers, hyphens
Scope of uniqueness Global (must be unique across all of Azure)
CAF pattern srch-{workload}-{environment}-{region}-{instance}
Worth knowing: Globally unique name.

Rules follow Microsoft's resource naming rules and CAF abbreviations. Some resource types have extra start/end character restrictions, so verify against the official docs before locking in a convention.

Naming AI + machine learning resources

Most AI and machine learning resources use the same Cognitive Services pattern: 2 to 64 characters, resource group scoped, not globally unique. That surprises people, since a lot of these accounts feel like they should behave like a database or storage account. In practice this means the region does not need to be in the name for uniqueness, but it is still worth keeping for clarity, since teams commonly provision the same service (OpenAI, Speech, Vision) in more than one region for quota or data residency reasons, and a name with no region becomes ambiguous the moment a second one exists.

Foundry hubs and projects are the one hierarchical pair in this category, a hub can contain several projects, so name the project for what it actually does and let the hub name carry the shared platform identity. A common mistake is naming multiple instances of the same service only by workload and skipping the purpose, Custom Vision, for example, has separate training and prediction resources that are easy to mix up once both exist side by side.

Example names

Generated by the same engine as the name generator, so length and character rules are already applied.

Scenario Generated name Length
Payments API, production srch-payments-prod-eus-001 26/60
Web app, development srch-webapp-dev-weu-002 23/60
Analytics with company prefix srch-contoso-analytics-stage-uks-001 36/60

Naming in Terraform and Bicep

Build the name from variables so one module produces the right name in every environment.

Terraform
locals {
  ai_search_name = lower("srch-${var.workload}-${var.environment}-${var.region}-001")
}
Bicep
var aiSearchName = toLower('srch-${workload}-${environment}-${region}-001')

Generate AI Search names for your workload

Free, no login. Validation and Terraform, Bicep, or CSV export included.

Open in the generator →

Related AI + machine learning resources

See the full list of Azure resource naming rules or read the complete CAF naming guide.