Edit

Foundry Agent Service feature availability in Azure Government

This article shows which Microsoft Foundry Agent Service agent types, tools, and publishing options are available in Azure Government — and which aren't. For platform-level features, endpoints, and portal URLs, see Microsoft Foundry in Azure Government. For an introduction to Agent Service, see the Foundry Agent Service overview.

Supported regions

Foundry Agent Service is available in the following Azure Government regions:

Region Region identifier
US Gov Virginia usgovvirginia
US Gov Arizona usgovarizona

Agent Service features and tools are available in both regions unless noted otherwise in the tables below.

Agent types

Agent type Available
Prompt agents Yes
Workflows Preview
Hosted agents No

Preview features in Azure Government

Features marked Preview are available for early adoption but might not carry the same compliance commitments (such as FedRAMP, DoD IL5, or CJIS) as generally available features. Confirm the compliance posture of any preview feature with your security and compliance team before using it for production or regulated workloads.

Tools

The following table lists tool availability for Foundry Agent Service in Azure Government. For details on each tool, see the Foundry tool catalog.

Tool Available
Code Interpreter Yes
Custom Code Interpreter Preview
File Search Yes
Azure AI Search Yes
Azure Functions Yes
Function calling Yes
Web search No
Grounding with Bing No
Image Generation No
Browser Automation No
Computer Use No
Microsoft Fabric No
SharePoint No
MCP servers No
Agent-to-Agent (A2A) No
OpenAPI tool No

Publishing

Azure Government supports publishing agents. Each published agent gets a stable managed endpoint and a Microsoft Entra identity. You can register the agent with the Entra Agent Registry for discovery within your tenant. Publishing to Microsoft Teams and Microsoft 365 Copilot isn't supported. For more information, see Publish an agent.

SDK configuration for Azure Government

Agent Service in Azure Government requires azure-ai-projects version 2.0.0 or later. Install the package with:

pip install "azure-ai-projects>=2.0.0" azure-identity

The token audience and project endpoint differ from public cloud. Set credential_scopes to https://ai.azure.us/.default and use your Azure Government project endpoint, which has the format https://{resource-name}.services.ai.azure.us/api/projects/{project-name}.

The following example creates a simple prompt agent, sends a message, and prints the response:

import os
from azure.identity import DefaultAzureCredential, get_bearer_token_provider
from azure.ai.projects import AIProjectClient

endpoint = os.environ["FOUNDRY_PROJECT_ENDPOINT"]
scope = "https://ai.azure.us/.default"

with (
    DefaultAzureCredential() as credential,
    AIProjectClient(
        endpoint=endpoint,
        credential=credential,
        credential_scopes=[scope],
    ) as project_client,
):
    api_key = get_bearer_token_provider(credential, scope)

    with project_client.get_openai_client(api_key=api_key) as openai_client:
        from azure.ai.projects.models import PromptAgentDefinition

        agent = project_client.agents.create_version(
            agent_name="MyAgent",
            definition=PromptAgentDefinition(
                model=os.environ["FOUNDRY_MODEL_NAME"],
                instructions="You are a helpful assistant",
            ),
        )
        print(f"Agent created (name: {agent.name}, version: {agent.version})")

        conversation = openai_client.conversations.create(
            items=[{"type": "message", "role": "user", "content": "Hello!"}],
        )

        response = openai_client.responses.create(
            conversation=conversation.id,
            extra_body={"agent_reference": {"name": agent.name, "type": "agent_reference"}},
        )
        print(f"Response: {response.output_text}")

        openai_client.conversations.delete(conversation_id=conversation.id)
        project_client.agents.delete_version(agent_name=agent.name, agent_version=agent.version)

For samples that use tools like file search, code interpreter, and function calling, see the agent SDK samples on GitHub.

Quotas, limits, and models