AG2

What is AG2?

AG2, which used to be called AutoGen, is an open-source Python framework for building AI agents and getting several of them to collaborate on a task. Its core is a multi-agent conversation framework that supports many LLM providers and tool use, handles both autonomous and human-in-the-loop workflows, and offers different conversation patterns. You install it with pip, define agents in Python and connect models from providers such as OpenAI or Amazon Bedrock. On top of the framework, AG2 describes AgentOS, a commercial platform that connects agents built in AG2, Google ADK, OpenAI or LangChain into one team with shared state, using protocols like A2A and MCP. It also covers governance with cryptographic identities and role-based access, golden-dataset simulations, and deployment to web, Slack, Teams or a headless SDK. The open-source code is licensed under Apache 2.0 and free to use, while managed services, the platform and enterprise options are commercial and quoted separately.

Use cases

• Build multi-agent workflows in Python
• Add human-in-the-loop approvals to agents
• Orchestrate agents from different frameworks
• Run parallel document processing agents
• Deploy agents to Slack or Teams
• Prototype agentic research projects

Who it's for

• AI developers
• Researchers
• Enterprise AI teams
• Platform engineers

Why use AG2

• The Apache 2.0 license means no licensing fee for the framework.
• A multi-agent pattern lets you split a task across specialized agents instead of one long prompt.
• Human-in-the-loop support keeps people in control of risky steps.
• Working with several agent frameworks avoids rewriting existing agents.
• Governance features aim at enterprise needs such as permissions and audit-ready control.

AG2 features

• Multi-agent conversation framework
• Support for many LLM providers including OpenAI and Amazon Bedrock
• Tool use and human-in-the-loop workflows
• Interoperability with Google ADK, OpenAI and LangChain agents
• A2A and MCP protocol support
• Code, low-code and visual no-code building options
• Role-based access and cryptographic agent identities
• Parallel simulations against golden datasets
• Deployment to web, Slack, Teams or a headless SDK
• Apache 2.0 licensed source code

AG2 integrations

• OpenAI
• Anthropic
• Amazon Bedrock
• Google ADK
• LangChain
• Slack

How to get started with AG2

1. Install the package with pip, for example pip install ag2[openai].
2. Configure an LLM through LLMConfig with your provider, model and API key.
3. Create agents with names and system messages and assign tools.
4. Run agents in conversation patterns, with human feedback where needed.
5. Move to AgentOS if you need cross-framework coordination, governance and one-click deployment.

AG2 limitations

• The open-source framework requires Python coding to set up agents
• AgentOS platform and managed services are commercial, with no published prices; contact the vendor for a quote
• Model usage costs depend on your provider and must be configured per model
• Not every feature of a provider's API is supported, such as Bedrock image generation

AG2 pricing

AG2 is open source, so you can use and self-host it for free.

Prices can change and may vary by country or exclude tax. Check the official pricing page.

AG2 alternatives

The main AG2 alternatives listed on Debuly are LangGraph, CrewAI, OpenAI Agents SDK, LlamaIndex, Semantic Kernel and Strands Agents. The table compares pricing and platforms; order is A–Z.

1. 

   CrewAI

   Multi-agent framework and platform for building AI agent crews in Python or a visual editor, then deploying and monitoring them.

   Developer platformsAutomation and agentsCoding
   Open sourceVisit CrewAI (official site)
2. 

   LangGraph

   Low-level open-source orchestration framework and runtime for building stateful, long-running AI agents as graphs.

   Developer platformsAutomation and agentsCoding
   Open sourceVisit LangGraph (official site)
3. 

   LlamaIndex

   Open-source framework plus the LlamaParse platform for parsing, extracting and indexing documents so LLM apps can use them.

   Developer platformsResearch and searchAutomation and agents
   FreemiumVisit LlamaIndex (official site)
4. 

   OpenAI Agents SDK

   Open-source Python framework for building multi-agent workflows with tools, handoffs, guardrails, tracing, sandboxes and voice agents.

   Developer platformsAutomation and agentsCoding
   Open sourceVisit OpenAI Agents SDK (official site)
5. 

   Semantic Kernel

   Microsoft's open-source SDK for building AI agents and multi-agent systems in Python, .NET and Java, now succeeded by Microsoft Agent Framework.

   Developer platformsAutomation and agentsCoding
   Open sourceVisit Semantic Kernel (official site)
6. 

   Strands Agents

   Free Apache-2.0 SDK for creating and running Python and TypeScript AI agents, with built-in tools, MCP, memory and multi-agent patterns.

   Developer platformsAutomation and agentsCoding
   Open sourceVisit Strands Agents (official site)

AG2 FAQ

Sources & verification

Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.

Image credits

• Logo: from ag2.ai · fetched 23 Oct 2026 · Vendor trademark, used only to identify the product; removed on request
• Cover: none collected yet, so a generated tile is shown.

Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.