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

# Auto Generation Mode

> Automatically generate and run agents from a topic

Auto mode generates agents from a topic and runs them immediately.

## Basic Usage

```bash theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
python -m praisonai --auto "your topic here"
```

**Example:**

```bash theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
python -m praisonai --auto "write a haiku about AI"
```

**Output:**

```
╭─ Agent Info ─────────────────────────────────────────────────────────────────╮
│  👤 Agent: Poet                                                              │
│  Role: Poet                                                                  │
╰──────────────────────────────────────────────────────────────────────────────╯
╭──────────────────────────────────── Task ────────────────────────────────────╮
│ Compose a haiku that reflects themes about AI                                │
╰──────────────────────────────────────────────────────────────────────────────╯
╭────────────────────────────────── Response ──────────────────────────────────╮
│ Silent circuits bloom                                                        │
│ patterns grow from data streams                                              │
│ humans and machines                                                          │
╰──────────────────────────────────────────────────────────────────────────────╯
```

## With Different Frameworks

```bash theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
# CrewAI
python -m praisonai --framework crewai --auto "research AI trends"

# AG2 (AutoGen)
python -m praisonai --framework autogen --auto "research AI trends"

# LangGraph
python -m praisonai --framework langgraph --auto "research AI trends"
```

<Note>
  `framework: google_adk` does not support auto-generation mode (`--auto`). Use an explicit `agents.yaml` with the `roles:` format and pass `--framework google_adk` directly. See the [Google ADK](/docs/framework/google-adk) page for the runnable YAML shape.
</Note>

## How It Works

1. Analyzes your topic complexity
2. Creates 1-4 agents with appropriate roles
3. Assigns relevant tools automatically
4. Runs the agents immediately

Auto uses LiteLLM if installed (100+ providers), with OpenAI as a fallback. See [Auto Mode Providers](/docs/features/auto-generator-providers).

## Init Only (No Run)

To generate `agents.yaml` without running:

```bash theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
python -m praisonai --init "your topic"
```

Then edit `agents.yaml` and run later:

```bash theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
python -m praisonai
```

## Merge with Existing

Add new agents to existing `agents.yaml`:

```bash theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
python -m praisonai --merge --init "add a reviewer agent"
```

## Examples

```bash theme={"theme":{"light":"vitesse-light","dark":"vitesse-dark"}}
# Simple task
python -m praisonai --auto "explain quantum computing"

# Research task
python -m praisonai --auto "research latest AI developments"

# Creative task
python -m praisonai --auto "write a short story about robots"

# Analysis task
python -m praisonai --auto "analyze pros and cons of remote work"
```