AU

AutoGen

Open-source framework for building agentic AI in Python and .NET

AutoGen is an open-source programming framework from Microsoft for building agentic AI systems. It gives developers code-level building blocks in Python and .NET to add agents to their own apps and services. Best suited for technical teams who want full control over how AI agents behave and run.

Open Source
iOS
API
Code Execution
Multi-Agent
B2B
For Developers
Visit AutoGen

Ask about AutoGen

Get answers based on AutoGen's actual documentation

Try asking:

About

What It Is

AutoGen is a GitHub-hosted programming framework for building "agentic AI"—AI agents that you embed and control from your own code. It’s aimed squarely at developers and technical teams who want to design, implement, and run agents inside existing applications rather than rely on a hosted assistant UI.

The project is structured as a multi-language SDK, with core libraries for Python and .NET and accompanying documentation. You install it into your codebase, configure it programmatically, and integrate it with your own infrastructure and AI model providers. The repository includes docs, devcontainer configuration, and CI pipelines, signaling that it’s intended for serious, production-style development workflows.

What to Know

AutoGen is a framework, not a turnkey assistant: you write code, wire up models and infrastructure, and decide how agents should behave. This makes it powerful and flexible for engineering teams, but inappropriate if you’re looking for a no-code or end-user-facing chatbot product. The GitHub repository clearly shows active development and versioned releases, but the snippet here doesn’t spell out which AI models or external services are supported, so you’ll need to consult the official docs for those details.

Because AutoGen is open source under MIT and CC-BY-4.0 licenses, you can run it entirely within your own environment, which is important if you need tight control over data and privacy. At the same time, you’re responsible for deployment, scaling, monitoring, and compliance; there’s no hosted SLA-backed service described in the available content. If you’re not comfortable managing code and infrastructure, or you need a plug-and-play SaaS agent, this framework is likely not the right fit.

Key Features
Open-source programming framework for building agentic AI systems
Language-specific SDKs for Python and .NET within a single unified project
Documentation site and in-repo docs to guide setup and integration
Repository structure geared toward production use, including Azure Pipelines CI configuration
Devcontainer configuration to simplify setting up a consistent development environment
Use Cases
Embedding custom AI agents into an existing Python or .NET application
Prototyping internal tools where autonomous agents handle repetitive or multi-step workflows under developer-defined rules
Building a backend service that exposes agentic AI capabilities via your own API or product UI
Agenticness Score
7/ 20
Level 1: Limited

Some tool use, but you're still driving

AutoGen is an open-source framework explicitly designed to build agentic AI into Python and .NET applications. Based on the provided description, it enables strong action capability through integration with existing services, data pipelines, and tools, and supports agentic workflows and orchestrations that imply moderate autonomy. It also exposes state-related APIs, suggesting that developers can manage and persist agent state across sessions. However, the supplied information does not describe any concrete mechanisms for error handling and adaptation, nor for safety, approvals, or monitoring. As a result, the overall agenticness profile, based only on the evidence given, is: strong action capability, moderate autonomy, basic statefulness, and no explicitly documented adaptation or safety features.

Total score: 3 (Action) + 2 (Autonomy) + 0 (Adaptation) + 2 (State) + 0 (Safety) = 7 → Level 1 (Limited agent, some tool use potential via integrations).

Score Breakdown

Action Capability
3/4
Autonomy
2/4
Adaptation
0/4
State & Memory
2/4
Safety
0/4

Categories

Pricing
  • Free / Open Source: Available under MIT and CC-BY-4.0 licenses.
  • Pro: Pricing not publicly available.
  • Enterprise: Pricing not publicly available.
Details
Website: github.com
Added: January 22, 2026
Last Verified: January 22, 2026
Agenticness: 7/20 (Level 1)
Cite This Listing
Name: AutoGen
URL: https://agentic-directory.onrender.com/t/autogen
Last Updated: January 29, 2026

Related Tools