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

# AI on mogenius

> Configurable AI agents, an interactive chat, and connectable models and tools — the AI capabilities built into your mogenius cluster.

mogenius brings AI directly into your Kubernetes operations. Instead of a single fixed assistant, you get a **toolbox of configurable agents**, an interactive **chat**, and a shared foundation of **models** and **tools** that both build on.

Everything runs against your own cluster through the mogenius operator and respects your existing roles and permissions.

## The building blocks

* **Agents** — Configurable AI workers that watch your cluster and produce troubleshooting reports and actionable proposals. You can run the curated defaults, adapt them, or create your own. See [Agents](/ai/agents).
* **Chat** — An interactive assistant for asking questions, investigating issues, and drafting manifests through natural language. See [Chat](/ai/chat).
* **AI Models** — One or more language models (OpenAI, Anthropic, or any OpenAI‑compatible / Ollama endpoint) that power agents and chat. See [AI Models & Setup](/ai/setup).
* **Tools** — Built‑in Kubernetes and Helm access, plus custom [MCP](https://modelcontextprotocol.io) servers you connect as additional agent tools. See [Tools & Custom MCP Servers](/ai/tools).

<Note>
  All AI features require at least one connected AI model. If you haven't set one up yet, start with [AI Models & Setup](/ai/setup).
</Note>

## Agents vs. Chat

Both use the same models and tools, but serve different jobs:

* **Agents** run **autonomously** — on a schedule, on cluster events, or on demand — and file their findings as **reports** you review and approve. Use them for continuous monitoring and remediation.
* **Chat** is **interactive** — you drive the conversation in real time. Use it for ad‑hoc investigation, questions, and quick edits.

## Where to find it

The AI features live under the **Agents** section in the platform:

* **Cluster level** — cluster‑wide agents and reports across every namespace the operator manages.
* **Workspace level** — the same view, scoped to the namespaces belonging to that workspace.

Models, tools, and custom MCP servers are managed centrally under **Cluster Settings → AI** and shared across the whole cluster.

## Get started

<Steps>
  <Step title="Connect a model">
    Add at least one AI model and mark a default. See [AI Models & Setup](/ai/setup).
  </Step>

  <Step title="Run or create an agent">
    Use a curated default agent or configure your own scope, triggers, and tools. See [Agents](/ai/agents).
  </Step>

  <Step title="Review reports">
    Approve, reject, or apply the proposals your agents generate. See [Reports & proposals](/ai/agents#reports-and-proposals).
  </Step>

  <Step title="Chat when you need it">
    Open the chat for interactive investigation. See [Chat](/ai/chat).
  </Step>
</Steps>