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

# Datum Assistant

> Ask Patch, the Datum Cloud assistant, about your resources in natural language, and let it diagnose problems for you.

Patch uses diagnostic tools from the Datum services your project uses, so it can work out why a Compute workload isn't available, why a load balancer isn't serving, or why a network isn't ready. It never changes anything on its own: it proposes a change as a plan, shows you the exact manifests, and applies them only when you approve.

Patch is available in the terminal with `datumctl assistant`. Conversations are kept by Datum, so you can pick one up from any machine you're signed in on.

```bash theme={null}
datumctl plugin install assistant
datumctl assistant chat "Why is the api-backend workload not available?"
```

Usage is metered by messages and tokens. See your organization's [Usage](https://datum.net/docs/organizations/usage.md).

> **[datumctl AI assistant](https://datum.net/docs/datumctl/ai/assistant.md)**
>
> Start a conversation, choose the project, and manage your conversations.

To connect your own AI tools to Datum instead, see [Datum MCP](https://datum.net/docs/agents/datum-mcp.md) and the [Agents overview](https://datum.net/docs/agents/overview.md).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.