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

# OpenGoat

> Build and run automated organizations of AI agents on top of OpenClaw.

OpenGoat is the easiest way to go from "one agent" to "a small AI team" that can plan work, delegate tasks, and keep context over time.

The first experience is intentionally simple:

1. Create a few agents
2. Message the Goat with your project idea
3. Watch the team come alive and start executing

## Quick Start

```bash theme={null}
npm i -g openclaw opengoat
openclaw onboard
opengoat start
```

Then open `http://127.0.0.1:19123`.

## Your dashboard is the starting point

<img src="https://mintcdn.com/opengoat/Uwq5nrQnp-X_1Ahe/images/dashboard.png?fit=max&auto=format&n=Uwq5nrQnp-X_1Ahe&q=85&s=081754d4ac842e3593c6b1fd3faa6d96" alt="Dashboard" width="3136" height="1348" data-path="images/dashboard.png" />

When you open the UI, you land on the Dashboard. It shows:

* **Organization Chart**: the hierarchical chart of your organization
* **Agent Sessions**: you can chat with any agent, or just interact with the Goat

You can always ask the Goat to "hire" new agents, but its recommended to create the first ones yourself.

## Step 1: Create a small team

On the Dashboard, click **Create Agent**, and select who the agent reports to.

#### Developers/Engineers

One key feature of OpenGoat, is that it support multiple agent runtimes.
You can run your developer agents on Claude Code, Codex, Cursor, etc.

## Step 2: Message the Goat to begin

In the left sidebar, under **Agents & Sessions**, click **Goat**.

The Goat will ask you about your project, will plan it, and break it down into smaller tasks.

## Step 3: Follow the work as it gets organized

Once the Goat starts, OpenGoat turns the plan into execution:

* tasks get created and assigned
* agents start working on their tasks one by one
* the Goat will delegate tasks accordingly to the organization size

You can watch things happen in:

* **Tasks**: see what’s open and what’s done
* **Agents**: inspect your org and who reports to whom
* **Logs**: view activity and runs

## What to do next

* [Docs Hubs](/start/hubs): task-oriented navigation for the full docs set
* [Quickstart](/start/quickstart): production path from install to first run
* [Installation](/start/installation): npm, Docker, and source installs
* [Organization Runtime](/core/organization-runtime): architecture and invariants
* [CLI Overview](/reference/cli-overview): command map and global flags