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

# Documentation

> Let investigations and the chatbot search your runbooks and reference docs.

Your team's documentation holds knowledge that lives nowhere else: the runbook for this exact failure, how a service is meant to behave, the decision behind an architecture. Connect your documentation and investigations can pull in the right runbook or reference at the right moment, and your team can search it directly from the chatbot.

## How investigations use this

Connected documentation is synced and indexed for both keyword and semantic search. From there it's available in two places:

* **In investigations**: relevant runbooks and reference docs surface as evidence, helping ground a finding in how your systems are actually meant to work and what to do about a known failure.
* **In the chatbot**: ask a question in an incident channel and the answer can cite the specific docs it drew on, with links back to the source.

<Info>
  Documentation is most valuable when it captures operational knowledge: runbooks, architecture overviews, and
  references. The more your docs describe how to operate and debug your systems, the more investigations can lean on
  them.
</Info>

## Providers you can connect

| Provider   | What gets synced                                |
| ---------- | ----------------------------------------------- |
| Confluence | Pages from the spaces you choose                |
| Notion     | Pages and databases you choose                  |
| GitHub     | Markdown and docs in repositories, by file path |
| GitLab     | Markdown and docs in repositories, by file path |

## Setup

Configure document sources from the [Investigations settings](https://app.incident.io/~/investigations/documents) in your dashboard.

<Steps>
  <Step title="Connect the provider">
    Connect the relevant integration (Confluence, Notion, GitHub, or GitLab) if you haven't already.
  </Step>

  <Step title="Choose what to sync">
    Scope each source to the docs that matter:

    * **Confluence**: select the spaces to sync.
    * **Notion**: select the pages or databases to sync.
    * **GitHub and GitLab**: choose repositories and the file paths to include (for example `docs/**` or `*.md`).
  </Step>

  <Step title="Let it sync">
    Documents are fetched, summarized, and indexed for search. Sources refresh automatically each day, and you can
    trigger a resync at any time.
  </Step>
</Steps>

<Note>
  When a document is removed at the source, it's dropped from search on the next sync. Only the documents you scope are
  synced: investigations never read beyond the spaces, pages, or paths you choose.
</Note>

## Related

<CardGroup cols={2}>
  <Card title="How investigations work" icon="magnifying-glass" href="/investigations/how-investigations-work">
    How a runbook becomes evidence in a finding.
  </Card>

  <Card title="Chatbot" icon="robot" href="/ai/at-incident">
    Ask questions in the incident channel and get cited answers.
  </Card>
</CardGroup>