Skip to main content
Microsoft Foundry is the current unified integration for Azure-hosted models in TrueFoundry’s AI Gateway. It uses one provider account for OpenAI models and models from publishers such as Anthropic, Cohere, Meta, and Mistral.

​
Microsoft Foundry

Use Microsoft Foundry for new Azure model accounts and migrations. It supports the OpenAI-compatible v1 surface as well as provider-specific routes required by Claude and Cohere deployments.
  • Configure the resource’s OpenAI endpoint and, for discovery and provider-specific models, its project endpoint.
  • Configure authentication once at the provider-account level.
  • Add integrations by Azure deployment name.
  • Use a Gateway model FQN such as my-foundry/gpt-4.1 in requests.
Existing Azure OpenAI and Azure AI Foundry resources remain available through their existing integrations. Microsoft Foundry is the unified path for new accounts and migrations.

​
Resource endpoints

Microsoft Foundry accounts can use two endpoint fields:
  • OpenAI endpoint: Used for the OpenAI-compatible /openai/v1 inference surface.
  • Project endpoint: Used for deployment discovery and resource-level provider routes such as Anthropic Messages and Cohere rerank.
For example:
Do not append /openai, /openai/v1, or an API path to the configured endpoint. The Gateway constructs the upstream path.

​
Existing Azure resources

Existing Azure OpenAI and Azure AI Foundry accounts continue to work with their current setup pages. Use Microsoft Foundry for new accounts that need one account across OpenAI and non-OpenAI deployments.

​
Next steps