> ## Documentation Index
> Fetch the complete documentation index at: https://docs.praison.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# AG2 with PraisonAI

> Use AG2 framework with PraisonAI for multi-agent GroupChat orchestration, tool registration, and AWS Bedrock support

AG2 is the community fork of AutoGen. PraisonAI supports AG2 as a dedicated framework backend with GroupChat orchestration, automatic tool registration, and native AWS Bedrock support.

<Note>
  Need a framework that isn't listed here? See [Framework Adapter Plugins](/docs/features/framework-adapter-plugins) to register your own via Python entry points.
</Note>

```mermaid theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
graph LR
    subgraph "AG2 Framework"
        YAML[📄 YAML Config] --> Agents[🤖 AG2 Agents]
        Agents --> GC[💬 GroupChat]
        GC --> Result[✅ Result]
    end

    classDef input fill:#6366F1,stroke:#7C90A0,color:#fff
    classDef process fill:#F59E0B,stroke:#7C90A0,color:#fff
    classDef output fill:#10B981,stroke:#7C90A0,color:#fff

    class YAML input
    class Agents,GC process
    class Result output
```

## Which AG2 Option to Choose

PraisonAI has two AG2-related framework options:

```mermaid theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
graph TB
    Start[Which option?] --> Q1{Need latest AG2<br/>features + Bedrock?}
    Q1 -->|Yes| A1["framework: ag2<br/>pip install praisonai[ag2]"]
    Q1 -->|No| Q2{Using legacy<br/>AutoGen v0.2/v0.4?}
    Q2 -->|Yes| A2["framework: autogen<br/>pip install praisonai[autogen]"]
    Q2 -->|New project| A1

    classDef question fill:#6366F1,stroke:#7C90A0,color:#fff
    classDef answer fill:#10B981,stroke:#7C90A0,color:#fff

    class Start,Q1,Q2 question
    class A1,A2 answer
```

| Option               | Install                            | When to Use                               |
| -------------------- | ---------------------------------- | ----------------------------------------- |
| `framework: ag2`     | `pip install "praisonai[ag2]"`     | New projects, Bedrock support, latest AG2 |
| `framework: autogen` | `pip install "praisonai[autogen]"` | Existing AutoGen v0.2/v0.4 projects       |

***

## Quick Start (AG2)

<Steps>
  <Step title="Install">
    ```bash theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
    pip install "praisonai[ag2]"
    ```
  </Step>

  <Step title="Create a YAML file">
    ```yaml theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
    framework: ag2
    topic: Research the latest developments in AI agents

    roles:
      researcher:
        role: AI Research Specialist
        goal: Find and summarize recent AI agent developments
        backstory: Expert researcher with deep knowledge of AI trends.
        tasks:
          research_task:
            description: Research the latest developments in {topic}
            expected_output: Summary report with key findings
    ```
  </Step>

  <Step title="Run">
    ```bash theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
    export OPENAI_API_KEY=your-key
    praisonai agents.yaml --framework ag2
    ```
  </Step>
</Steps>

***

## How AG2 Works

```mermaid theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
sequenceDiagram
    participant PraisonAI
    participant UserProxy as UserProxy Agent
    participant GroupChat
    participant Agents as AG2 Agents

    PraisonAI->>UserProxy: Create with TERMINATE check
    PraisonAI->>Agents: Create from YAML roles
    PraisonAI->>GroupChat: Create with all agents
    UserProxy->>GroupChat: Initiate chat
    GroupChat->>Agents: Route messages
    Agents->>Agents: Collaborate
    Agents-->>UserProxy: TERMINATE
    UserProxy-->>PraisonAI: Final result
```

AG2 uses a **GroupChat** pattern: a `UserProxy` agent initiates the conversation, and a `GroupChatManager` routes messages between your defined agents. Agents collaborate until one says "TERMINATE".

***

## Multi-Agent Example

```yaml theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
framework: ag2
topic: Write a blog post about renewable energy

roles:
  researcher:
    role: Energy Research Analyst
    goal: Gather facts about renewable energy trends
    backstory: Expert in energy research and data analysis.
    tasks:
      research_task:
        description: Research current renewable energy trends for {topic}
        expected_output: Research brief with statistics and trends

  writer:
    role: Content Writer
    goal: Write engaging blog content based on research
    backstory: Skilled writer who turns technical research into readable content.
    tasks:
      writing_task:
        description: Write a blog post based on the research findings
        expected_output: Complete blog post with introduction, body, and conclusion
```

```bash theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
praisonai agents.yaml --framework ag2
```

***

## AWS Bedrock Support

AG2 natively supports AWS Bedrock models. No API key needed - uses your AWS credentials.

```yaml theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
framework: ag2

roles:
  cloud_architect:
    role: Cloud Solutions Architect
    goal: Design cloud architectures
    backstory: Expert in AWS cloud infrastructure.
    llm:
      model: bedrock/anthropic.claude-3-5-sonnet-20241022-v2:0
      api_type: bedrock
      aws_region: us-east-1
    tasks:
      design_task:
        description: Design a serverless architecture for a web app
        expected_output: Architecture diagram and component descriptions
```

Bedrock credentials come from your standard AWS setup (`~/.aws/credentials`, environment variables, or IAM role).

***

## Auto Mode

```bash theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
praisonai --framework ag2 --auto "Create a Movie Script About Cat in Mars"
```

***

## Legacy AutoGen (v0.2 / v0.4)

For existing projects using the older AutoGen package:

```bash theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
pip install "praisonai[autogen]"
```

```yaml theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
framework: autogen
topic: Create Movie Script About Cat in Mars

roles:
  researcher:
    role: Research Analyst
    goal: Gather information about Mars and cats
    backstory: Skilled in research, with a focus on gathering accurate and relevant information.
    tasks:
      research_task:
        description: Research about Mars environment and cat behavior
        expected_output: Research findings document with key facts
```

```bash theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
praisonai agents.yaml --framework autogen
```

### AutoGen Version Selection

```bash theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
# Use AutoGen v0.4 (default if available)
export AUTOGEN_VERSION=v0.4

# Use AutoGen v0.2
export AUTOGEN_VERSION=v0.2

# Auto-select (default: prefers v0.4)
export AUTOGEN_VERSION=auto
```

***

## Framework Selection Priority

1. **CLI flag** (`--framework ag2`) takes precedence
2. **YAML file** (`framework: ag2`) is used if no CLI flag
3. **Default**: praisonai framework

<Note>
  The `roles` format YAML is required for both `ag2` and `autogen` frameworks. The newer `steps` + `agents` workflow format only supports the praisonai framework.
</Note>

<Warning>
  Direct prompts always use the praisonai framework regardless of the `--framework` flag:

  ```bash theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
  # This will use praisonai, NOT ag2
  praisonai "What is 2+2?" --framework ag2
  ```

  To use AG2, provide a YAML file with the `roles` format.
</Warning>

## Troubleshooting

**Unsupported framework error**: If you see `Unsupported framework: autogen. Registered: ['ag2', 'autogen', 'autogen_v4', 'crewai', 'praisonai']`, check spelling — frameworks are registered via the adapter registry.

**Missing installation errors**:

* `AG2 is not available. Install with: pip install ag2` — Install AG2 dependencies
* `AutoGen v0.2 is not available. Install with: pip install autogen` — Install AutoGen v0.2 dependencies

***

## Best Practices

<AccordionGroup>
  <Accordion title="Use AG2 for new projects">
    The dedicated `ag2` framework uses the latest AG2 package with `LLMConfig` support and native Bedrock. Use `autogen` only for existing codebases.
  </Accordion>

  <Accordion title="Keep agent roles focused">
    Each role should have a clear, distinct responsibility. The GroupChat manager routes messages based on role descriptions.
  </Accordion>

  <Accordion title="Use Bedrock for AWS environments">
    Set `api_type: bedrock` in the role's `llm` config. No API keys needed when running on AWS with proper IAM roles.
  </Accordion>
</AccordionGroup>

***

## Related

<CardGroup cols={2}>
  <Card title="CrewAI" icon="users" href="/framework/crewai">
    CrewAI framework integration
  </Card>

  <Card title="Agents" icon="user" href="/concepts/agents">
    PraisonAI native agents
  </Card>
</CardGroup>