microsoft/Agents

The Microsoft 365 Agent SDK simplifies building full stack, multichannel, trusted agents for platforms including M365, Teams, Copilot Studio, and Webchat.

What it solves

It simplifies the process of building and deploying AI agents across various communication channels, such as Microsoft 365 Copilot, Microsoft Teams, and custom web apps. It removes the restriction of being tied to a specific AI technology stack, allowing developers to use their preferred AI services and models to power their agent's logic.

How it works

The SDK provides a standardized "container" for agents that handles state, storage, and the management of activities and events. This container acts as a bridge between the AI logic and the deployment channel. It is available in C#, JavaScript, and Python, and is designed to be agnostic of the AI provider, meaning developers can integrate any LLM or orchestration framework (like Semantic Kernel) to drive the agent's behavior.

Who it’s for

Developers who want to build enterprise-grade AI agents that can be deployed to Microsoft 365 ecosystems or custom applications without being locked into a specific AI vendor or language.

Highlights

  • Channel Agnostic: Deploy the same agent logic to Microsoft Teams, Microsoft 365 Copilot, and other custom apps.
  • AI Agnostic: Works with any AI service, model, or technology stack.
  • Multi-language Support: Dedicated client libraries for C#/.NET, JavaScript, and Python.
  • Developer Tooling: Includes AI coding assistant plugins for GitHub Copilot and Claude Code to help with SDK implementation and debugging.

Related

  • Project
    CopilotKit/channels-sdk

    An SDK that connects AI agents to Slack, Microsoft Teams, and Discord, enabling native interactive UI and human-in-the-loop approvals.

  • Project
    github/copilot-sdk

    A set of multi-language SDKs that let developers embed Copilot's agentic workflows and runtime into their own applications.

  • Project
    openai/openai-agents-python

    A lightweight, provider-agnostic framework for building multi-agent workflows, supporting text, voice, and sandboxed execution across various LLMs.

  • Project
    strands-agents/harness-sdk

    An open-source SDK for Python and TypeScript that provides a model-driven agent loop and infrastructure for building AI agents without a hosted control plane.

  • Project
    lofcz/LLMTornado

    A .NET provider-agnostic SDK for building and orchestrating AI agents and workflows with built-in connectors to 30+ API providers and vector databases.