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

# What's Ordyn

> Ordyn is the AI production engineer. It watches every change you ship, keeps your alerting tuned, and fixes production when it breaks.

## How it works

> **Detect**
>
> Ordyn reads every pull request and creates the monitors that change needs. After deploy, it confirms the change is healthy and traces regressions back to the PR.

> **Triage**
>
> Every alert, from Ordyn or from tools you already have, is investigated and graded before it reaches Slack. Only real impact gets through.

> **Fix**
>
> What lands in Slack comes with the root cause, the evidence, and a fix. Ordyn can open the PR itself or hand off to your coding agent.

## What Ordyn needs

* **GitHub**, so Ordyn can read your code, review pull requests, and open fixes.
* **Slack**, so Ordyn can deliver results where your team already works.
* **At least one observability source**, such as Sentry, Datadog, or your cloud logs. More sources mean better investigations.

Ordyn's access to your observability tools and databases is read-only. The only things Ordyn writes are the monitors it creates, its pull request reviews and fixes, and its Slack messages. Your data is never used to train models. See [Security](/reference/security).

## Get started

> **[Quickstart](/quickstart)**
>
> Connect your stack and turn on Auto Monitors in about 15 minutes.

> **[Shadow mode](/shadow-mode)**
>
> Run Ordyn next to your real alert channel for a week before changing anything.

> **[Integrations](/integrations/overview)**
>
> Everything Ordyn connects to and how to set it up.

> **[Investigate on demand](/slack/investigate-on-demand)**
>
> Ask Ordyn about your logs, metrics, or code any time.


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