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

# Fern

> Import published Fern documentation into AGO, including Brevo developer guides and API references

Connect a public Fern documentation site to make its guides and API reference pages available to your AGO agents. No API key or GitHub access is required.

## Before you start

You need access to **Knowledge → Sources** and a public documentation site that exposes an `llms.txt` index and Markdown pages. Password-protected documentation is not supported.

## Connect your documentation

<Steps>
  <Step title="Create a source">
    Open **Knowledge → Sources**, create a source, give it a name, and select **fern** as its type.
  </Step>

  <Step title="Enter the documentation URL">
    Enter the site URL in **Documentation URL**. For Brevo, use `https://developers.brevo.com`.

    To import just one section, enter its section URL, provided it has an `llms.txt` index. Use the site URL rather than an individual article such as `/docs/getting-started`.
  </Step>

  <Step title="Choose the pages to include">
    Leave both pattern lists empty to import every page listed in the index, including nested indexes. Otherwise, add the path patterns you want to include or exclude.
  </Step>

  <Step title="Save and sync">
    Save the source, then start a sync. Follow progress in **Sync history**. Assign the source to an agent to use the imported documentation in answers.
  </Step>
</Steps>

## Filter pages

Patterns match the full URL path, starting with `/`, without the `.md` extension. `*` matches any sequence of characters. Exclusions take precedence over inclusions.

| Setting          | Example                 | Result                               |
| ---------------- | ----------------------- | ------------------------------------ |
| Include Patterns | `/docs/*`               | Import Brevo pages under `/docs/`    |
| Include Patterns | `/docs/getting-started` | Import only the getting-started page |
| Exclude Patterns | `/changelog/*`          | Skip changelog entries               |

Pages keep their titles and links to the original website. They appear as a flat list in index discovery order; Fern's sidebar hierarchy is not reproduced.

AGO removes recognized default Fern instructions about Markdown exports, the documentation index, and MCP setup when they appear at the very beginning of a page. Customized wording or unsupported formatting may remain. The same text inside the page body or a code example is preserved.

Later successful syncs update existing pages and remove pages no longer listed or included by your filters. Manual content overrides are preserved. If a fetch fails or no pages match, the sync fails without removing old documents; pages updated before the failure remain updated.

## Troubleshooting

* **No pages found:** Check the site or section URL and your filters. Open its `/llms.txt` URL to confirm it lists documentation pages.
* **HTML instead of Markdown:** Check that the site is public and the page's `.md` URL returns content rather than a sign-in screen.
* **Redirect outside the site:** Enter the final documentation site URL. Imports stay on the configured scheme, host, and section path.
* **Large imports:** Narrow the URL or use include patterns. Each response is limited to 10 MB, and discovery supports up to 1,000 indexes and 20,000 selected pages. Linked OpenAPI files and other downloads are excluded.

If your documentation is available only as a checked-out repository on the server, see [File Directory](./file-directory-integration). The Fern connector reads published content and does not connect directly to GitHub.