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

# Introduction

> apowerb is an open-source agentic framework for building, running and operating AI agents.

apowerb is an agentic framework for building and deploying AI agents. It ships as a
Python package exposing a FastAPI server, a Typer CLI, and a REST API that any client
can drive — including the companion Next.js interface, [apowerb-ui](/ecosystem/apowerb-ui).

Agents are defined in the database, materialised as Python modules at startup, and
executed through [Google ADK](https://google.github.io/adk-docs/). Model access goes
through [LiteLLM](https://docs.litellm.ai/), so any major provider works without
touching agent code.

<CardGroup cols={2}>
  <Card title="Quickstart" icon="rocket" href="/quickstart">
    Bring the full stack up with Docker in a few minutes.
  </Card>

  <Card title="Install from PyPI" icon="python" href="/installation">
    Run the server on its own, against your own Postgres.
  </Card>

  <Card title="API reference" icon="code" href="/api-reference/introduction">
    Every endpoint, generated from the OpenAPI schema.
  </Card>

  <Card title="Contributing" icon="code-pull-request" href="/contributing/contributing">
    How to propose a change and get it released.
  </Card>
</CardGroup>

## What it does

| Capability              | Summary                                                                                            |
| ----------------------- | -------------------------------------------------------------------------------------------------- |
| **Agent orchestration** | Four execution patterns — `base`, `parallel`, `sequential`, `loop` — plus hierarchical sub-agents. |
| **Tool system**         | 24+ tool modules: Google Workspace, Microsoft 365, databases, RAG, S3, visualisation, web search.  |
| **RAG as a service**    | Index files, URLs, SQL query results and S3 buckets into per-agent knowledge bases.                |
| **Text-to-SQL**         | Schema introspection plus natural-language to SQL, executed against your database.                 |
| **Webhooks**            | Gmail (Pub/Sub) and Outlook (Graph) push notifications trigger agents on new mail.                 |
| **OAuth integrations**  | Google Workspace, Microsoft 365 and GitHub connected as agent tools.                               |
| **SSE streaming**       | Token-by-token agent output, RAG indexing progress, and user notifications.                        |
| **Scheduling**          | Cron-style recurring agent runs.                                                                   |
| **Agent Hub**           | Publish an agent, let another workspace clone it.                                                  |

## Requirements

* Python 3.12 or later (3.13 recommended)
* PostgreSQL
* [uv](https://docs.astral.sh/uv/) for local development

<Note>
  Some capabilities described across this documentation belong to commercial editions
  and are **absent** from the open-source build — not disabled behind a flag. See
  [Editions](/concepts/editions) for how to tell which is which.
</Note>