> For the complete documentation index, see [llms.txt](https://docs.katara.ai/about/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.katara.ai/about/agents-and-answering/integrations.md).

# Existing integrations

Integration records may remain visible even when there is no current creation journey. Use this guide to maintain an existing configuration, not as a promise that every connector can be installed from the dashboard.

## Before you start

Select the organization that owns the agent and confirm your current edit and integration permissions. Have the existing agent identified before changing external channel connections. Review [Availability](https://docs.katara.ai/about/reference/availability.md) for deployment limitations.

**Create an agent** currently offers only **Content Uploader** and **Q\&A**. Brain creation is separate. Website, GitHub, and Notion source-ingestion agents are retired: do not use old creation links or assume their retained forms mean configuration updates or execution are accepted.

## Inspect an existing integration

1. Open **Agent Library**, search for the existing agent, and open its details.
2. Review its configuration and any connected **Workflows** before changing it. These per-agent relationships are not a Workflows index in the library.
3. For existing Discord, Telegram, or Slack agents, inspect **Servers**, **Groups**, or **Channels**, respectively. Use only the controls currently available to your account.
4. For an existing Website Widget, open **Embed Widget** when authorized. This surface requires permission to create organization tokens; review configuration and embedding settings before copying code into a website.
5. Reopen the relevant configuration after saving and verify the connected relationship. Test the actual external channel or website separately before relying on it for users.

Existing screens do not certify live provider credentials or successful delivery. If installation or saving fails, preserve the error and ask the workspace operator to check availability and authority. Do not recreate a retired source agent as a workaround.

## Google Drive: incomplete direct-route workflow

The direct `/dashboard/apps-integrations` route renders Google Drive controls, but it is not linked from the primary sidebar and is not a recommended onboarding path.

If your operator directs you there, use **Connect Google Drive**, review connection status, then **Browse files** to search, page through files, select, and queue them. **Disconnect** is available for the connection. The visible queue is local page state; persisted queue loading and a deterministic OAuth callback handshake remain incomplete. A queued row is not proof of durable ingestion, and returning to the page may not restore it.

For a standard file-ingestion path, use [Content Uploader](https://docs.katara.ai/about/agents-and-answering/content-uploader.md) or [Brain sources](https://docs.katara.ai/about/brains/sources.md).

## Keep GitHub workflows distinct

Brain GitHub export publishes committed Wiki history through Brain settings. It is not the retired GitHub ingestion agent, does not import repository files, and is not bidirectional synchronization. Begin with [Brains](https://docs.katara.ai/about/brains/overview.md) for that separate workflow.
