Skip to content
ReviewRamble

The weekly brief

AI, reviewed, in your inbox.

  • New reviews of tools, models, MCP servers and repos, with the verdict up front
  • Re-checked prices and ratings when something changes
  • New analysis pieces as they are published

Models, agents, MCP servers and the tools actually worth your time.

One email a week. Unsubscribe with the link in any brief. How we use your email

AutoGen

Microsoft's influential multi-agent framework, now archived with no bug or security fixes; Microsoft Agent Framework replaces it.

Reviewed by ReviewRamble editors · Updated

View on GitHub Open source

Share this page:XLinkedInFacebookRedditWhatsAppEmail

Our rating

2.6/5

AutoGen is archived and unmaintained, so it belongs in legacy code you are migrating and in the history of agent frameworks, not in a new project.

What is AutoGen?

AutoGen is Microsoft Research's Python and .NET framework for multi-agent applications. It has three layers: a Core API for event-driven agents and local or distributed runtimes, AgentChat for common patterns such as two-agent chat and group chat, and Extensions for model clients, code execution and MCP tools. AutoGen Studio adds a no-code interface for prototyping, AutoGen Bench a benchmarking suite, and Magentic-One is a reference multi-agent team built on it.

The README now carries an archived banner: Microsoft provides no bug fixes, security fixes, new features or support, advises against production use, and points new users to Microsoft Agent Framework and its AutoGen migration guide. The last release, python-v0.7.5, came out on 30 September 2025. The last code change, on 26 March 2026, restricted dynamic provider loading to trusted namespaces; later commits only touch documentation and archive notices.

Is AutoGen worth it?

AutoGen's status is the review: Microsoft has stopped maintaining it. The README says there will be no bug fixes, security fixes, new features or support, that vulnerabilities in AutoGen and its dependencies may remain unaddressed, and that it was built for research and experimentation and is not advised for production. Its contributing note still describes a maintenance mode limited to bug fixes, security patches and documentation, but since the last code fix on 26 March 2026 every commit has changed only docs and archive banners. The repository is not locked, and community help continues on Discord and GitHub Discussions, where Microsoft warns that responses may be limited.

It earned its 61,923 stars. AutoGen made conversational multi-agent patterns popular, with group chats where agents take turns, and a layered design in which AgentChat sits on a Core runtime that can run across processes and between Python and .NET. The Extensions layer added MCP workbenches, so an agent can drive a Playwright browser through an MCP server in a few lines; the README itself warns to connect only trusted MCP servers, since they may run commands on your machine.

The catch: the risks compound over time. AutoGen Studio, the drag-and-drop interface many people met first, is described in its own README as a prototype that is not meant for production, leaving authentication and security to you, which matters for a tool whose agents can execute code. 1,181 issues and pull requests are open, and Microsoft will not fix them. The code is MIT licensed, so a fork is legal, but whoever forks it owns every future patch.

Our take: Microsoft Agent Framework is the direct successor, built by the same teams, with stable 1.0 APIs and an official AutoGen migration guide; for Python teams outside Azure, LangGraph or CrewAI are the more common landing spots. AutoGen is worth reading for its ideas and worth running only inside a migration plan.

Where it earns its place

  • A platform engineer with AutoGen 0.7 agents in production

    Plan a safe exit now that Microsoft has stopped patching it

    Pin the current packages, list which AgentChat patterns you rely on, and follow Microsoft's AutoGen to Agent Framework migration guide. Until the port is done, any new vulnerability in AutoGen or its dependencies is yours to fix.

  • A university lecturer teaching multi-agent systems

    Show students how group chat and agent-as-tool orchestration work

    The AgentChat examples run in a few lines on Python 3.10 or newer with an OpenAI key. Run them in a sandbox, since agents may execute code; the only cost is model tokens.

  • A researcher reproducing a published experiment built on AutoGen

    Re-run the original agent setup to compare against a new method

    Install the exact package versions the paper used from PyPI in an isolated environment. AutoGen Bench, the project's benchmarking suite, is still in the repository.

Before you commit

Setup
pip install -U "autogen-agentchat" "autogen-ext[openai]" on Python 3.10 or newer; AutoGen Studio installs with pip install -U autogenstudio and serves its UI on localhost:8080.
Integrations
OpenAI and Azure OpenAI model clients through autogen-ext, MCP servers through McpWorkbench, code execution extensions, and .NET packages on NuGet.
Data and privacy
Runs in your own process; prompts and data go to whichever model and MCP servers you configure. Microsoft warns that unaddressed vulnerabilities raise the risk of handling sensitive data with it.
Cost at scale
Free under MIT; the real cost is model tokens plus the engineering time to patch a fork or migrate to a maintained framework.
Limits
No release since September 2025, no Microsoft support, limited community responses, and AutoGen Studio is not meant for production.

What’s good

  • Clear layered design: Core runtime, AgentChat and Extensions
  • Two-agent and group-chat patterns that shaped later frameworks
  • MIT-licensed code, so a fork is possible
  • Python and .NET packages with a cross-language runtime

What’s not

  • No bug fixes, security fixes or support from Microsoft
  • Last release, python-v0.7.5, in September 2025
  • AutoGen Studio is a prototype, not a production app
  • 1,181 open issues and pull requests

Best for

Teams keeping existing AutoGen code alive while they migrate, and researchers studying multi-agent patterns.

AutoGen alternatives

AutoGen: common questions

Is AutoGen worth it?
AutoGen is archived and unmaintained, so it belongs in legacy code you are migrating and in the history of agent frameworks, not in a new project. We rate it 2.6 out of 5.
How much does AutoGen cost?
Free and open source: code under the MIT licence, documentation under CC BY 4.0. You pay only for the models you call.
What are the downsides of AutoGen?
No bug fixes, security fixes or support from Microsoft; last release, python-v0.7.5, in September 2025; autoGen Studio is a prototype, not a production app; 1,181 open issues and pull requests.
Who is AutoGen best for?
Teams keeping existing AutoGen code alive while they migrate, and researchers studying multi-agent patterns.
How much does AutoGen cost for a team or in production?
Free under MIT; the real cost is model tokens plus the engineering time to patch a fork or migrate to a maintained framework.
Is AutoGen safe for company data?
Runs in your own process; prompts and data go to whichever model and MCP servers you configure. Microsoft warns that unaddressed vulnerabilities raise the risk of handling sensitive data with it.
What are the best AutoGen alternatives?
The closest alternatives we have reviewed are Microsoft Agent Framework, LangGraph, CrewAI and OpenAI Agents SDK.

Answers checked Nov 18, 2026.