agent-framework-azure-ai

What it is and what it does

This package extends Microsoft Agent Framework with Azure AI Foundry memory integration, allowing agents to maintain semantic memories of user profiles and conversation context. It provides a FoundryMemoryProvider that automatically retrieves static memories on first interaction, searches for contextual memories based on ongoing conversations, and updates the memory store with new messages.

The package wraps several Azure AI dependencies (azure-ai-projects, azure-ai-agents, azure-ai-inference) and the core agent-framework components to create a turnkey memory layer. It is currently in beta (1.0.0rc6) and supports Python 3.10 through 3.14. The integration is designed for developers building conversational agents that need persistent, searchable memory without manual state management.

Use it for:

• Building customer support agents that remember user preferences and past interactions across multiple sessions
• Creating personalized conversational AI that learns and recalls user context automatically during conversations
• Implementing multi-turn dialogue systems where agents need to reference earlier conversation topics without explicit prompting
• Developing agents that maintain user profiles and adapt behavior based on remembered preferences
• Prototyping agent memory features using Azure's managed memory infrastructure without building custom persistence

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Integrates Azure AI Foundry memory capabilities into Microsoft Agent Framework, enabling agents to store, retrieve, and maintain semantic memories across conversations.

Yes, if you are building agents on Microsoft Agent Framework and need semantic memory backed by Azure. The package is actively maintained, has low install friction, and integrates cleanly with the broader Azure AI ecosystem. The beta status (1.0.0rc6) means the API may change before 1.0 stable, so pin the version in production. No known security vulnerabilities.

Install

agent-framework-azure-ai on PyPI

pip install agent-framework-azure-ai

uv add agent-framework-azure-ai

poetry add agent-framework-azure-ai

Installing agent-framework-azure-ai

Before you install

Low friction install with a pure Python wheel. Active maintenance with recent commits and a substantial repository. Still in beta (1.0.0rc6), so expect API surface changes before a stable release.

License in practice

MIT license permits commercial and private use with minimal restrictions, making it suitable for most projects.

Quickstart

pip install agent-framework-azure-ai --pre

from agent_framework_azure_ai import FoundryMemoryProvider
from azure_ai_projects import AIProjectClient

# Create memory store and provider
memory_provider = FoundryMemoryProvider(project_client)
# Attach to agent for automatic memory retrieval and updates

Requires Python 3.10 or later and valid Azure credentials configured for AIProjectClient.

Verify before relying

• Whether the memory store automatically persists across application restarts or requires explicit save/load logic
• Performance characteristics when retrieving contextual memories from large conversation histories
• Whether configurable memory update delays can be tuned for real-time vs. batch scenarios

Package facts

Evidence: agent_framework_azure_ai-1.0.0rc6-py3-none-any.whl

Tags

More Application Frameworks packages

Further reading