# Microsoft Agent Framework — Build & Run Agent Workflows

> Microsoft Agent Framework builds production-grade agents and multi-agent workflows in Python and .NET, with official docs and quickstarts for teams.

## Install

Save as a script file and run:

## Quick Use

1. Install / run:
   ```bash
   pip install agent-framework

dotnet add package Microsoft.Agents.AI
   ```
2. Start / smoke test:
   ```bash
   python -m pip show agent-framework && echo 'MAF installed (Python)'
# .NET: verify package in your project
# dotnet list package | rg Microsoft.Agents.AI
   ```
3. Verify:
   - Follow the official Quick Start link from the README and run the smallest end-to-end sample before integrating into your app.

## Intro

Microsoft Agent Framework builds production-grade agents and multi-agent workflows in Python and .NET, with official docs and quickstarts for teams.

- **Best for:** Teams building production agents who want official quickstarts, multi-language support, and a structured workflow model
- **Works with:** Python + .NET (per README); integrates with official docs and tutorials
- **Setup time:** 15 minutes

## How to Use It Well

- Setup time ~15 minutes (pip/dotnet install + follow the Quick Start tutorial)
- Two supported stacks called out in the README: Python and .NET
- GitHub stars + forks (verified): see Source & Thanks

Treat agent frameworks like application frameworks: start from an official quickstart, lock in your model/tool boundaries, then add evals, telemetry, and permission controls before shipping.

### FAQ

**Q: Is it only for prototypes?**
A: The README positions it for production-grade agents and workflows; still validate your own safety and eval gates.

**Q: Can I use only Python?**
A: Yes. The README shows Python installation via pip; .NET is optional depending on your stack.

**Q: Where do I start?**
A: Use the official documentation and Quick Start linked in the README, then adapt the sample pattern to your domain.

## Source & Thanks

> Source: https://github.com/microsoft/agent-framework
> License: MIT
> GitHub stars: 10,345 · forks: 1,692

---

<!-- ZH -->

## 快速使用

1. 安装 / 运行：
   ```bash
   pip install agent-framework

dotnet add package Microsoft.Agents.AI
   ```
2. 启动 / 冒烟测试：
   ```bash
   python -m pip show agent-framework && echo 'MAF installed (Python)'
# .NET: verify package in your project
# dotnet list package | rg Microsoft.Agents.AI
   ```
3. 验证：
   - Follow the official Quick Start link from the README and run the smallest end-to-end sample before integrating into your app.

## 简介

Microsoft Agent Framework（MAF）提供 Python 与 .NET 的多语言框架，用于构建生产级 AI agent 与多 Agent 工作流，并配有官方 quick start 与文档。

- **适合谁：** 想用官方 quick start + 多语言支持构建生产级 agent/多 agent 工作流的团队
- **可搭配：** Python + .NET（见 README）；可结合官方文档与教程使用
- **准备时间：** 15 分钟

## 实战建议

- 上手约 15 分钟（安装 + 跑通 Quick Start 教程）
- README 明确支持两条技术栈：Python 与 .NET
- GitHub stars + forks（已核验）：见「来源与感谢」

把 agent 框架当应用框架来用：先跑通官方 quickstart，明确模型/工具边界，再把评测、可观测性与权限控制补齐后再上线。

### FAQ

**它只适合原型吗？**
A: README 将其定位为生产级框架；但仍建议你自己加安全与评测门槛再上线。

**我只用 Python 可以吗？**
A: 可以。README 提供了 pip 安装路径；.NET 取决于你的技术栈。

**从哪里开始最省事？**
A: 直接从 README 里的官方文档与 Quick Start 入手，先跑通最小样例再迁移到业务。

## 来源与感谢

> Source: https://github.com/microsoft/agent-framework
> License: MIT
> GitHub stars: 10,345 · forks: 1,692


---
Source: https://tokrepo.com/en/workflows/microsoft-agent-framework-build-run-agent-workflows
Author: Agent Toolkit